{ ; ¢ \ i | i I ; i $ 3 i y \' re ri : " } Vy i Fogarty International Center Series on PREVENTIVE MEDICINE Volume 4 { DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE 4 Public Health Service National Institutes of Health I = ee DIABETES MELLITUS The John E. Fogarty International Center for Advanced Study in the Health Sciences National Institutes of Health Bethesda, Maryland 1976 Stefan S. Fajans, M.D. Editor DHEW Publication No. (NIH) 76-854 DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE Public Health Service National Institutes of Health For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402 RC 660 U454L PUBL ORGANIZING COMMITTEE AND EDITORIAL BOARD Stefan S. Fajans, M.D., Chairman Peter H. Bennett, M.D. Norbert Freinkel, M.D. David Kipnis, M.D. Paul E. Lacy, M.D. Albert I. Winegrad, M.D. iii -] Ef Ur Eo nap BT rg re x] oi A od A i LES [hu k ow og EES PREFACE The Fogarty International Center was established in 1968 as a memorial to the late Congressman John E. Fogarty from Rhode Island. It had been Mr. Fogarty's desire to create within the National Institutes of Health a center for research in biology and medicine dedicated to international cooperation and collaboration in the interest of the health of mankind. The Fogarty International Center is a unique resource within the Federal establishment, providing a base for expansion of America's health research and health care to lands abroad and for bringing the talents and resources of other nations to bear upon the many and varied health problems of the United States. As an institution for advanced study, the Fogarty International Center has embraced the major themes of medical education, environmental health, societal factors influencing health and disease, geographic health problems, interna- tional health research and education, and preventive medicine. Our commitment to the study of preventive aspects of human disease is expressed in the forth- coming Fogarty International Center Series on Preventive Medicine. Improvement in the health status of the American people will depend, in great measure, on the design and application of programs which place major emphasis on the preventive aspects of human disease. Although health authori- ties generally agree with this thesis, there is need for more precise definition of effective methods and programs of prevention, financial resources required to implement these programs, and priorities to be assigned to research in pre- ventive methodology. The need to assemble expertise in this field to elucidate mechanisms whereby the full impact of preventive medicine may be brought to bear on the solution of America's major health problems has been expressed repeatedly in public statements by leaders throughout the health field. In response to this need, the Fogarty International Center initiated a series of comprehensive studies of preventive medicine in order to review and evaluate the state of the art of prevention and control of human diseases, to identify the deficiencies in knowledge requiring further research, including analysis of financial resources, preventive techniques, and manpower, and to recognize problems in application of preventive methods and suggest corrective action. This monograph, Diabetes, is the product of diligent effort by numerous experts in the field whose contributions have been blended into a single volume by a skillful editorial committee under the chairmanship of Dr. Stefan S. Fajans and represents the fourth volume of the Fogarty International Center Series on Preventive Medicine. The subtle nature of diabetes mellitus and the broad spec- trum of the clinical expressions of this disease has made precise definitions and determinations of prevalence difficult. Conservative estimates reveal that vi approximately 400,000 Americans each year learn for the first time that they have diabetes, four million are presently affected and 40,000 are said to succumb to one of the complications of this disease. Ranking as the sixth commonest cause of mortality and accounting for the expenditure of at least two billions of dollars, diabetes mellitus is clearly a major health problem in this country. This monograph presents a detailed review and evaluation of the state of knowledge and research in the prevention and control of diabetes mellitus, identifies gaps in the knowledge of this disease for the purpose of defining research needs, and proposes way in which current knowledge may be applied more effectively to alleviate the burden of diabetes. Milo D. Leavitt, Jr., M.D. Director Fogarty International Center CONTRIBUTORS Sol Aisenberg, Ph.D. Space Science Division, Whittaker Corporation, Boston, Massachusetts Ronald A. Arkey, M.D. Division of Medicine, Mount Auburn Hospital, Cambridge, Massachusetts Walter F. Ballinger II, M.D. Department of Surgery, Washington University School of Medicine, St. Louis, Missouri Bernard Becker, M.D. Professor of Opthalmology, Washington University School of Medicine, St. Louts, Missouri Peter Bennett, M.D. } Epidemiology and Field Studies Branch, National Institute of Arthritis, Metabolism, and Digestive Diseases, Phoenix, Arizona Robert E. Bolinger, M.D. Department of Internal Medicine, University of Kansas Medical Center, Kansas City, Kansas Ronald A. Chez, M.D. Pregnancy Research Branch, National Institute of Child Health and Human Development, National Institute of Health, Bethesda, Maryland K.W. Chang, Ph.D. Space Science Division, Whittaker Corporation, Boston, Massachusetts Rex S.. Clements, Jr., M.D. Division of Endocrinology and Metabolism, Department of Medicine, Uni- versity of Alabama Medical Center, Birmingham, Alabama R. H. Egdahl, M.D. Department of Surgery, University Hospital, Boston, Massachusetts Paul S. Entmacher, M.D. Medical Director, Metropolitan Life Insurance Company, New York City, New York Frederick H. Epstein, M.D. Professor of Epidemiology, University of Michigan, Ann Arbor, Michigan Stefan S. Fajans, M.D. Professor of Internal Medicine, University of Michigan Medical Center, Ann Arbor, Michigan F. Robert Fekety, Jr., M.D. Professor of Internal Medicine, University of Michigan Medical Center, Ann Arbor, Michigan James B. Field, M.D. Clinical Research Unit, Department of Medicine, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania Daniel W. Foster, M.D. Department of Internal Medicine, Southwestern Medical School, The University of Texas Health Science Center at Dallas, Texas Norbert Freinkel, M.D. Kettering Professor of Internal Medicine, Northwestern University, Chicago, Illinois Frederick C. Goetz, M.D. Division of Internal Medicine, University of Minnesota Medical School, Minneapolis, Minnesota Jack Goldstein, B.A., B.M.E. University of Southern California School of Medicine, Los Angeles, Cali- fornia J. M. Hiebert, M.D. Department of Surgery, University Hospital, Boston, Massachusetts vii viii Ronald Kalkhoff, M.D. Milwaukee County General Hospital, The Medical College of Wisconsin, Mil- waukee, Wisconsin R. C. Karl, M.D. Washington University School of Medicine, St. Louts, Missouri David Kipnis, M.D. g . . Professor of Internal Medicine, Washington University Medical School, Barnes and Wohl Hospital, St. Louis, Missouri Harvey C. Knowles, Jr., M.D. Department of Internal Medicine, University of Cincinnati Medical Center, Cineinnati, Ohio Robert A. Kreisberg, M.D. Department of Medicine, University of South Alabama, Mobile, Alabama Paul Lacy, M.D. Department of Pathology, Washington University School of Medicine, St. Louis, Missouri Bernard Landau, M.D. Department of Medicine, Case Western Reserve University, Lakeside Hospital, Cleveland, Ohio Arnold Lazarow, M.D. Department of Anatomy, University of Minnesota Medical School, Minneapolis, Minnesota Harold E. Lebovitz, M.D. Division of Endocrinology, Duke University Medical Center, Durham, North Carolina Rachmiel Levine, M.D. City of Hope National Medical Center, Duarte, California K. Lundbaek, Professor AARHUS Universitet, Kommunehospitalet, AARHUS, Denmark Fred R. McCrumb, Jr., M.D. Fogarty International Center, National Institutes of Health, Bethesda, Maryland Curtis L. Meinert, Ph.D. University of Maryland School of Medicine, Division of Clinical Investiga- tion, Baltimore, Maryland Leona V. Miller, M.D. University of Southern Califormia Medical Center, Diabetes Section, Los Angeles, California Daniel H. Mintz, M.D. Department of Medicine, University of Miami School of Medicine, Miami, Florida John A. Moorhouse, M.D. Faculty of Medicine, The University of Manitoba, Winnipeg General Hospital, Winnipeg, Manitoba, Canada Bryce D. Munger, M.D. Department of Anatomy, College of Medicine, The Pennsylvania State Uni- versity, The Milton S. Hersey Medical Center, Hershey, Pennsylvania Ruth Osterby, M.D. Patologist-Anatomisk Institut, Kommunehospitalet, AARHUS, Denmark Leon D. Ostrander, M.D. Professor of Internal Medicine, University of Michigan, University Hospital, Ann Arbor, Michigan Daniel Porte, Jr., M.D. Veterans Administration Hospital, Seattle, Washington Thaddeus Prout, M.D. Greater Baltimore Medical Center, Baltimore, Maryland Arthur H. Rubenstein, M.D. Department of Medicine, University of Chicago Hospitals and Clinics, Chicago, Illinois John W. Runyan, Jr., M.D. Division of Health Care Sciences, University of Tennessee, Memphis, Tennessee D. W. Scharp, M.D. Washington University School of Medicine, St. Louis, Missouri Joseph Silva, M.D. g ae ; Assistant Professor of Internal Medicine, University of Michigan Medical School, Ann Arbor, Michigan Morton E. Smith, M.D. 3 Assoctate Professor of Opthalmology, Washington University School of Medicine, St. Louis, Missouri J. Stewart Soeldner, M.D. Associate Professor of Medicine, Harvard University Medical School, Joslin Research Laboratory, Boston, Massachusetts D. F. Steiner, M.D. Biochemistry Department, University of Chicago Hospitals and Clinics, Chicago, Illinois Albert Winegrad, M.D. George S. Cox Medical Research Institute, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania ix = iE - by " =F oo ps ." LT NEE ® hoa 5 ay ohh pls } “Byer, I il 3 ve 4 Ae! FR A by gi lia Fi CONTENTS Organizing Committee and Editorial Board Preface Contributors Contents Introduction Chapter 1 The Problem of Diabetes Mellitus Stefan S. Fajans and Norbert Freinkel .....eeeeecessssscscss EPIDEMIOLOGY Chapter 2 Diabetes Mellitus: The Overall Problem and Its Impact on the Public Harvey C. Knowles, Jr., Curtis L. Meinert, and Thaddeus E. Prout ceeesceecvee ER eT El al TT Chapter 3 Economic Impact of Diabetes Paul S. ENCMACNGY sseevrvinsnonso suse viosnsnsinee saness vanes Chapter 4 Improving the Organization of Care for the Chronically Ill Leona V. Miller, Jack Goldstein, and John W. Runyan, Jr. «eeesees cso siemens nsnsee vee sessemsese wees Chapter 5 The Computer in the Management of Diabetes Robert E. Bolinger ...eceeseee “esessesavsssessesecnsres esevacs ETIOLOGY Chapter 6 Islet Cell Dysfunction A. H. Rubenstein and D. F. BLeindr ..vesesssnssssvsrnsinssvns Chapter 7 Infectious and Immune Mechanisms in the Etiology and/or Pathogenesis of Diabetes Mellitus Bryce L. Munger ...... 3 Ee Siete wid re wearers terol iP re wieriiie ele sin FACTORS INFLUENCING DEVELOPMENT OF THE DIABETIC STATE Chapter 8 Environmental Factors Influencing the Development of the Diabetic State Ronald A. ArkY css cevescvesvsvecsnsnsnns Wve sie vie eis susie swe cece Chapter 9 The Role of the Neuroendocrine System in the Development of Diabetes Mellitus Daniel Porte; JLe sesvesvssorsnsee eee ACUTE COMPLICATIONS OF THE DIABETIC STATE Chapter 10 Diabetic Coma Daniel W. FOSter ...eceeees cress SSR NA Wi Se Fie vis .v Page iii vii xi xiii 11 33 41 47 59 73 89 106 123 xi Chapter 11 Hyperosmolar Coma ” Chapter 13 James B., Field .oeveeese lL TTY Cr inivinaies NE EE Se lactic Acidosis: Interrelationships with Diabetes Mellitus and Phenformin Robert A. KrelSherg «.evssncvsssnis won ae Eee See Sis avin See Acute Complications of the Diabetic State Joseph Silva and F. RODErt FeRely ..ecessvsesinnonsvsnse sions LONG-TERM COMPLICATIONS + Chapter + Chapter Chapter Chapter Chapter Chapter 14 15 16 17 18 19 Demonstrable Metabolic Abnormalities in Diabetes Mellitus that May Contribute to the Pathogenesis of Specific Late Complications A. I. Winegrad and Rex 'S. Clements, JL. ssvsvivavsinirssnsis Diabetes, Hyperglycemia and Atherosclerosis: New Research Directions Leon D. Ostrander, Jr. and Frederick H. Epstein ...eeeeeess. Ocular Complications in Diabetes Morton -E. Smith and Bernard BeCKeY ..seessssssavevsnvsie verse Renal Disease in Diabetes Mellitus K. Lundbaek and R. @sterby ....... REP TUIPE TIN. SNRPCIC A IS CN Diabetic Peripheral Neuropathy J. A. MOOThOUSE eveevvevnnnnenn ais eam ame eee AW Hie wi anein sie sin new Effect of Diabetes Mellitus on Fetal Growth and Development Daniel H. Mintz and Ronald A. Ch€Z ....cevs. a ve sees say NEW APPROACHES TO CONTROL AND PREVENTION Chapter Chapter 20 21 Chapter 2 Chapter 23 Chapter 24 xii Chapter 25 Diabetes Mellitus: A Bioengineering Approach: An Implantable Glucose Sensor J. S. Soeldner, K. W. Chang, Sol Aisenberg, J. M. Hiebert, and R. H., EGAN] woes cinsninininosevves avons Transplantation of Insulin Secreting Tissues R. C. Karl, D. W. Scharp, Paul E. Lacy, and W. FP. BALIINGOT «vovvrrnvsvsvvssvosnsnvsenisssssssionssnns Diet and Diabetes Mellitus Ronald KR. RBIKHOEE vive vsvanamevinesvensvis gn yes weston sone Insulin Synthesis and Analogs HAPOLd Zui LODOVIEZ + os wv vis vvin slusiv vi sit sim mn in wid vin sivas sreidisle vio 4 Drugs Enhancing Insulin Secretion Harold E. LeDOVIEZ civiieniveiensvsinomvivsiosnaivonmse sens vines Drug Altering Carbohydrate and Lipid Metabolism Bernard Robert Landau ...... o sie vw ten sine Hie ww saws sie. eesvissieve 133 142 154 173 194 213 227 243 256 267 278 295 310 327 344 INTRODUCTION EE EER SR Diabetes mellitus has become a major public health problem when con- sidered on the basis of its incidence, morbidity, mortality, and socio- economic impact on the world's populations. The establishment of clear-cut objectives to guide a broad program of prevention and treatment of the disease and its complications is of primary importance. Within the framework of the Fogarty International Center Series on Pre- ventive Medicine, a six-member Organizing Committee was established to guide a review of problems in the field of diabetes mellitus. The charge to the Committee was to prepare a monograph which would (1) review and evaluate the present state of knowledge and research in the basic and the clinical aspects of diabetes mellitus; (2) assess the applicability of present knowledge to the prevention and control of diabetes, and (3) identify gaps in knowledge and areas requiring further research. The Committee asked contributors to the monograph to (1) make projections and recommendations for future needs in terms of manpower and techniques, and (2) consider the direction in which maximum efforts might be most fruitful in developing measures or means for the preven- tion of clinical diabetes and its complications. It is obvious that effective preventive measures are difficult, if not impossible, where the etiology and pathogenesis are poorly understood. Never- theless it was visualized that a cooperative venture representing several points of view might serve as a stimulus for accelerative research to gain new basic knowledge. The monograph would not serve as a textbook of diabetes mellitus but as a document prepared primarily for the use of professional and/or administrative personnel and planners of the National Institutes of Health, as well as a resource for legislative committees. It was the hope of the Committee that the report would be a valuable contribution to national health planning and would assist in the making of informed decisions about mea- sures that might have significant impact in terms of public health. Accord- ingly, the Organizing Committee intended each chapter to include material which would give (1) background, (2) the current state of knowledge, (3) a realistic appraisal of what information must still be acquired through research now and in the future, and (4) how and why such information should lead to im- proved preventive medicine (a) immediately by application to patient care and (b) by long-range projection. It is apparent that some subject matters lend themselves better to this approach than do others. The table of contents details the contributions and their authors. The various sections address themselves to the problem of the definition of diabetes, epidemiology, socioeconomic aspects of the disease, etiology and pathogenesis, acute complications of the diabetic state, long-term complications, new approaches to control and prevention of the disease, and manpower needs for research and patient care. The Committee hopes that the monograph will fulfill the needs which prompt- ed its formulation. The Committee is very much indebted to Dr. Fred R. McCrumb of the Fogarty International Center for his assistance in all aspects of the planning and preparation. Organizing Committee and Editorial Board November 1975 xiii = a 7 pk . ie i pe ul” hp Oe — = ia fom : 5 DIABETES MELLITUS ped hid . i ety i pe i » = th " PE oo 3 es efit, THE PROELEM OF DIABETES MELLITUS Stefan S. Fajans and Norbert Freinkel "Idiopathic' diabetes mellitus' is a disorder of metabolism that, in its fully developed clinical expression, is characterized by fasting hyperglycemia, atherosclerotic and microangio- pathic vascular disease and neuropathy. Hyperglycemia may become manifest years before the clinical recognition of vascular disease or neuropathy. Only a few decades ago a generally ac- ceptable definition of diabetes mellitus would have stressed the presence of continuous hyperglycemia and glycosuria. Even today a few investigators and clinicians are hesitant to ac- cept a definite diagnosis of diabetes in the absence of fasting hyperglycemia (i.e., continuous hyperglycemia). However, the great majority agree that diabetes mellitus may present clinically in a mild or asymptomatic form without fasting hyperglycemia and that this is the most common recognizable form of the disease. The typical vascular and neuropathic manifestations of diabetes may occur in patients with relatively mild carbohydrate intolerance and with normal fasting blood glucose levels. A definition of ''carbohydrate intolerance' is difficult since there is no strict separation between normal and abnormal carbohydrate tolerance even in otherwise healthy, ambula- tory young people. It is based on statistical rather than biological considerations. Even in most groups of first degree relatives of diabetic patients there is no sharp dividing line but a continuum between normal and abnormal carbohydrate metabolism. There is common agreement that "idiopathic'" diabetes is a disease in which an inherited susceptibility may play an important part. This susceptibility has its origin at conception and may exist for prolonged periods before additional pathogenetic factors cause the emergence of a recognizable abnormality of carbohydrate metabolism. The genetic defect may remain without clini- cal expression indefinitely. Thus, a definition of genetic diabetes mellitus should include stages in the natural history of the disease which presently cannot be recognized, since we lack a marker for ''genetic diabetes." Before discussing further some arbitrary definitions of the types and stages in the natural history of '"idiopathic' diabetes, it is important to cite several recent findings and concepts. The biochemical and clinical manifestations of the disease encompass a spectrum from the unrecognizable, to the recognizable but asymptomatic form of the disease, and to symptomatic diabetes with acute metabolic decompensation (ketoacidosis, hyperosmolar coma) or with chronic complications or associations (cataracts, complications of pregnancy, neuropathy, atherosclerosis, microangiopathy). Hyperglycemia and these complications are found in both juvenile-onset type and maturity-onset type diabetes. These two types of diabetes usually have been thought to represent only a quantitative difference in the defect in insulin secretion or action, and attendant se- quellae. However, a growing body of evidence suggests the existence of heterogeneity of '"idio- pathic diabetes mellitus" in terms of (a) inheritance, (b) insulin responses to glucose in maturity-onset type diabetes, and (c) prevalence of vascular disease. lAs distinguished from secondary types of diabetes or carbohydrate intolerance associated with a variety of well-defined genetic syndromes. 2 Diabetes Mellitus A difference in the inheritance of diabetes has been shown between the families of maturity- onset type diabetes in young people and families of patients with classical juvenile-onset type diabetes (1). This difference not only provides evidence of genetic heterogeneity but further indicates that there is a need for careful definition of the phenotype of diabetes in popula- tions in which the genetics of diabetes is to be analyzed. Genetic heterogeneity has been found also among sets of identical twins of which at least one had diabetes mellitus (2). Concordance of diabetes among the pairs of identical twins was very high (92 percent) among those in whom the age of onset of diabetes in the index twin was 40 years or more (mostly maturity-onset type), while concordance was found with a frequency of only 53 percent in those in whom diabetes was diagnosed under 40 years of age in one twin (mostly juvenile-onset type). This suggests that there is a difference in genetic as well as environmen- tal factors in the etiology and pathogenesis of diabetes between these two groups of identical twins. A difference in the inheritance between juvenile-onset and maturity-onset type diabetes may be associated with a difference in the frequency of occurrence of certain histocompatibility types or HL-A antigens (HL-A8 and/or W15) (3,4), and a difference in the frequency with which viral and autoimmune processes may be involved (3,4). In one group of patients with juvenile-onset type diabetes an increased susceptibility to beta cell damage by viral agents may be due to a defective immune response influenced by genes in the HL-A chromosomal region and leading to an autoimmune process (3,4). The presence of cell-mediated immunity to pancreas antigen was reported to be more frequent in patients with insulin-dependent than in patients with insulin-independent diabetes (3, 5). These findings support the concept that these two types of diabetes differ from each other and have been cited to indicate also that they are two different disease entities in etiology and pathogenesis (3). In all prevalence figures based on population statistics, diabetes mellitus has been treated as a single disease entity. Juvenile-onset type, ketosis-prone diabetes can be associated with all the acute and chronic complications of the disease. Since it is associated with absolute in- sulin insufficiency and with the sudden appearance of typical symptomatology, it is usually cited as the prototype of diabetes. On the other hand, among the total population of known diabetic patients not more than 5 percent belong to this type, while the great majority can be character- ized as maturity-onset type or ketosis-resistant diabetic patients. The definitions of the stages in the natural history of diabetes mellitus have been based on the absence or presence and on the degree of abnormality of glucose metabolism (6). These have ranged from prediabetes or potential diabetes, where there is no recognizable abnormality of car- bohydrate metabolism, to overt or clinical diabetes mellitus with gross fasting hyperglycemia. Prediabetes or potential diabetes is a conceptual state which can be recognized only in retrospect since we lack a specific marker for ''genetic' diabetes and since there is heterogeneity of dia- betes in terms of genetics. Thus, at the present time, we cannot determine what percentage of a general population harbors a genetic predisposition to diabetes providing the necessary base for environmental factors to precipitate the disease. On the other hand, latent or chemical diabetes is the stage in which a diagnosis of diabetes can be made by the use of standardized laboratory procedures although the patient is completely asymptomatic. The majority of patients with maturity-onset type diabetes without complications The Problem 3 have asymptomatic, latent, or chemical diabetes. Heterogeneity of insulin responses to administered glucose among nonobese patients with latent diabetes has been demonstrated and emphasizes that so-called "idiopathic'' diabetes mellitus includes more than one disorder associated with hyperglycemia (6,7). Progression to insulin- requiring diabetes (some to ketosis-prone type) occurred only in individuals who had insulin re- sponses which were delayed and subnormal, or lower than the mean responses of the control sub- jects. In contrast, none of the patients with 'high'" insulin responses, or with responses ex- ceeding the mean of the control subjects, have progressed to insulin-requiring diabetes. Heterogeneity in occurrence of significant vascular disease in diabetes also suggests that we are dealing with a syndrome that includes entities in which different pathogenetic factors (genetic, environmental) are at play. Among a group of classical, juvenile-onset type, ketosis- prone diabetics of more than 30 years duration and with continuous hyperglycemia ("poorly con- trolled") 20 percent of patients do not have clinically significant retinopathy or nephropathy (8). "As reported from at least six clinics in five different countries, clinical evidence of microvas- cular disease, atherosclerosis, or neuropathy has been found in only 20-40 percent of insulin- requiring, ketosis-prone patients, who have survived diabetes for 20-40 years or more (9). These concepts of the heterogeneity of "idiopathic diabetes mellitus' in terms of heredity, insulin responses to glucose, and occurrence of vascular disease underline the enormous task which lies ahead of us in trying to unravel the etiologies and pathogeneses of diabetes mellitus and its complications. The plural is used advisedly since it is becoming apparent that we are dealing not with a single disease but with a syndrome composed of multiple entities. However, although genetic etiologies may differ, certain uniform metabolic consequences can be anticipated whenever blood glucose remains elevated too much and too long. For example, as re- viewed elsewhere in this volume (Chapter 13, Winegrad and Clements), hyperglycemia effects an in- creased disposition of glucose by the polyol pathway. Such diversion of glucose could alter intracellular ion fluxes, hydration, reductive environment, and other metabolic parameters as yet undefined. Hyperglycemia can also induce selective increase of enzymes involved in forming the constituents of basement membranes (Chapter 13, Winegrad and Clements). Even more subtle physio- chemical changes may result from hyperglycemia: e.g., serum osmolality appears to parallel hyperglycemia in asymptomatic maturity-onset diabetics (10). Presumably, therefore, thirst is not activated sufficiently to compensate for the osmotic contribution of the '"extra' circulating glucose in this relatively stable population. Since all cells should function as perfect osmom- eters, one may infer that intracellular hyperosmolarity is a common concomitant of hyperglycemia and that the hydration of intracellular macromolecules may be compromised more frequently in dia- betics than is generally appreciated (10). Regardless of the underlying genetic defect(s), any or all of the above consequences of hyperglycemia per se could contribute to the development of some of the anatomic lesions of diabetes. In addition, certain untoward aspects of the diabetic state could be caused by secondary in- creases in the availability of hormones with potentially catabolic properties. For example, exu- berant basal levels of growth hormone are more common in diabetics and paradoxical responses of growth hormone to glucose occur more frequently (see Chapter 13, Winegrad and Clements). Growing clinical experience suggests that these disturbances in growth hormone secretion may parallel the severity of the metabolic derangements rather than being characteristic of any single subgroup of 4 Diabetes Mellitus diabetics (11,12). Indeed, the abnormalities may be rectified in large measure by regulation of the diabetes (12). Failure of glucose to effect normal suppression of glucagon has also been re- ported in diabetic subjects (13). It is possible that this altered pattern of glucagon secretion may represent a primary component of genetic diabetes especially since the alpha and beta cells share a common embryological origin. Moreover, as of this writing, complete correction of gluca- gon dysfunction has not yet been observed in human diabetics at least during acute insulin ther- apy (13). However some secondary changes in glucagon secretion can be induced. Thus, the same pattern of seeming glucagon autonomy has been elicited in animals with experimental diabetes, and, under such circumstances, it.can be reversed by treatment with insulin (14). Clearly, definitive answers concerning primary vs. secondary etiologies for abnormalities in the secretion of hormones other than insulin in genetic human diabetes must await experiences with insulinization programs that reproduce the normal multiphasic excursions of endogenous insulin more closely than conven- tional therapy (e.g., the "artificial pancreas'; islet transplants, etc.; see Chapters 19 and 20). However, on the basis of existing data, it seems likely that at least some of the disturbances in hormones such as growth hormone and glucagon are secondary to acquired changes in glucose recogni- tion within the cells from which these hormones are secreted, or within those suprahypophyseal or peripheral neural centers which regulate their responsiveness to metabolic mixtures. All the above represent pathophysiological phenomena which could be ascribed to hyperglycemia and disturbed glucose homeostasis per se. As such, they convey the tacit implication that simple correction of the abnormalities in blood sugar (and the relative insulinopenia) should also elim- inate many of the complications of diabetes mellitus. However, despite a voluminous and con- troversial literature, that proposition has never been really tested. As cited above, none of our traditional forms of diabetes therapy have reduplicated the moment-to-moment fine tuning between circulating glucose and insulin that the normal beta cell provides. Moreover, it may be simplis- tic to view hyperglycemia in such unidimensional terms. Immediately following alimentation, metabolic objectives are geared to utilize ingested nutrients not only for fulfillment of pre- vailing oxidative demands and repair of antecedent metabolic deficits, but also for storage in anticipation of the longer intervals when exogenous fuels will be unavailable. With carbohydrate- containing "mixed meals," the prompt and appropriate acute increases in the availability of insu- lin, and the counterbalancing reductions in hormones such as glucagon, and usually growth hormone, make optimum anabolism possible (15). But the anabolism involves dietary protein and fat as well as carbohydrate, and in an integrated and insulin-dependent fashion. Thus, it is perhaps mis- leading to speak of 'glucose intolerance,'" even though historical precedent and our diagnostic use of simple glucose challenges to reproduce the 'fed state" (15) have made this term part of common clinical parlance. Since insulin affects anabolism of and from protein and fat, as well as carbohydrate, a logical extrapolation from ''glucose intolerance' must be that the complex inte- grated disposition of all dietary components is also "out of phase.” Supporting data are sparse, and may require greater analytical sophistication and more diagnostic studies with "mixed meals" rather than oral or intravenous glucose. It is not inconceivable that dose-response curves be- tween insulin and individual food-stuff may differ so that given degrees of ''glucose intolerance" may be accompanied by varying delays in the concurrent disposition of the amino acids, chylomi- crons, and very low density lipoproteins of dietary origin. Similarly, 'fasting hyperglycemia" cannot be viewed in terms of glucose alone. Normoglycemia after overnight fast presupposes that The Problem 5 basal elaboration of insulin has been sufficient to brake the recall of endogenous fuels in accord with prevailing oxidative demands. ''Fasting hyperglycemia" therefore indicates that cata- bolism has been inappropriate, and by definition, such faulty catabolism involves stores of pro- teins and lipids, as well as carbohydrates. In this regard, the changes in blood glucose that occur with aging warrant comment. Epide- milogical experience has indicated that fasting blood glucose increases at the rate of "2 mg/100 ml per decade through the adult years;' "postprandial values change at the rate of 4 mg/100 ml per decade and those following a glucose challenge at 6-13 mg/100 ml per decade" (16,17). It is not yet certain whether this is a normal consequence of aging or because increased age unmasks those with a genetic predisposition to one of the forms of diabetes mellitus. Suffice it to say that within the framework of 'mormal criteria," some 50-60 percent of all sexugenarians and septugenar- ians may display abnormal ''glucose tolerance" (17,18). Moreover, population studies have confirmed that hypertension and ECG abnormalities are more common in those with such asymptomatic elevations in post glucose values (see Chapter 14). Within the context of the metabolic considerations that are outlined above, one may infer that sluggish anabolism from ingested nutrients and less regu- lated catabolism of endogenous resources may obtain in fully half of the aged (17,18). It is tempting to speculate that some of the so-called 'mormal' concomitants of old age, such as atherosclerosis, may represent expressions of these metabolic abnormalities; and that in this regard, some of the lesions of the young and middle-aged diabetic represent a form of "accelerated aging." The observation that the regression coefficient between age and width of the basement membrane of muscle capillaries is twice as great in diabetics as in normals is compatible with this proposition (19). Thus, the apparent sparing of some diabetics from degenerative complica- tions (8) might merely reflect a genetic resistance to the metabolic consequences of faulty post- prandial anabolism and/or poorly integrated catabolism--i.e., a genetic propensity to longevity. In the least however, the metabolic view of all diabetes as a form of "accelerated aging" would imply that detailed analysis of the diabetic syndromes may yield not only new genetic insights but also more catholic clues concerning those degenerative changes which we have classically accepted as "age-related." In Chapters 2 and 3 the prevalence of diabetes in terms of known hyperglycemia, the preva- lence of macroangiopathy, microangiopathy, and neuropathy in those with diagnosed diabetes, as well as the social and economic aspects of the disease are given. Population statistics have rarely been obtained by examining populations particularly predisposed to the disease. It has been reported that in the families of young propositi with maturity-onset type diabetes, 85 per- cent of the parents and 53 percent of siblings had maturity-onset type diabetes as well (1). Furthermore, it has been shown that in the majority of patients with this type of diabetes there may be no, or only slow, progression of carbohydrate intolerance over many years or decades and that diabetes may remain unrecognized for similar periods of time (6). Several population studies have demonstrated an impressive correlation between the prevalence of asymptomatic hyperglycemia and the prevalence of atherosclerotic cardiovascular disease (20-22). We do not know in what pro- portion of patients with cardiovascular disease the disease is secondary to or associated with diabetes mellitus. We are also lacking large scale longitudinal prospective studies of individ- uals with asymptomatic diabetes in early life to determine the frequency of emergence of overt diabetes or of complications such as neuropathy, atherosclerotic cardiovascular disease, 6 Diabetes Mellitus retinopathy, or nephropathy in middle or advanced age. Such studies are essential if we are to" assess the magnitude of the problem which diabetes poses in terms of morbidity and mortality and if we are to devise effective prophylactic procedures in the future. REFERENCES 1. Tattersall, RB, and SS Fajans 1975. A difference between the inheritance of classical juvenile-onset and maturity-onset type diabetes of young people. Diabetes 24:44-53. 2. Tattersall, RB, and DA Pyke 1972. Diabetes in identical twins. The Lancet 2:1120-1125. 3. Nerup, J, P Platz, 00 Andersen, M Christy, J Lyngsge, JE Poulsen, LP Ryder, M Thomsen, LS Nielsen, and A Svejgaard 1974. HL-A antigens and diabetes mellitus. The Lancet 2:864-866. 4. Cudworth, AG, and JC Woodrow 1974. HL-A antigens and diabetes mellitus. The Lancet 2:1153. 5. MacCuish, AC, J Jordan, CJ Campbell, LJP Duncan, and WJ Irvine 1974. Cell-mediated immunity to human pancreas in diabetes mellitus. Diabetes 23:693-697. 6. Fajans, SS, CI Taylor, JC Floyd, Jr, and JW Conn 1974. Some aspects of the natural history of Diabetes Mellitus. Excerpta Medica Int'l Cong Series No 312, pp 329-340. 7. Fajans, SS, JC Floyd, Jr, CI Taylor, and S Pek 1975. Heterogeneity of insulin responses in latent diabetes. Trans Assoc Amer Phys 87:83-94. 8. Knowles, H 1971. Long-term juvenile diabetes treated with unmeasured diet. Trans Assoc Amer Phys 85:95-101. 9. Oakley, WG, DA Pyke, RB Tattersall, and PJ Watkins 1974. Long-term diabetes. A clinical study of 92 patients after 40 years. Quart J Med 43:145-156, (Jan). 10. Singer, DL, ME Drolette, D Hurwitz, and N Freinkel 1962. Serum osmolality and glucose in maturity onset diabetes mellitus. Arch Intern Med 110:758-762. 11. Molnar, GD, WF Taylor, A Langworthy, and V Fatourechi 1972. Diurnal growth hormone and glucose abnormalities in unstable diabetics: Studies of ambulatory-fed subjects during continuous blood glucose analysis. J Clin Endocrinol 34:837-846. 12. Corrall, RJ, WM Hunter, IW Campbell, ADB Harrower, LJP Duncan, and BF Clarke 1974. Reversal by insulin treatment of abnormal growth hormone pattern in newly diagnosed diabetes mellitus. Acta Endocrinol 77:115-121. 13. Unger, RH, LL Madison, and WA Meller 1972. Abnormal alpha cell function in diabetics-- response to insulin. Diabetes 21:301-307. 14. Mueller, WA, GR Faloona, and RH Unger 1971. The effect of experimental insulin deficiency on glucagon secretion. J Clin Invest 50:1992-1999. 15. Freinkel, N, and BE Metzger 1969. Oral glucose tolerance curve and hypoglycemias in the fed state. N Eng J Med 280:820-828. 16. O'Sullivan, JB 1974. Age gradient in blood glucose levels: Magnitude and clinical implica- tions. Diabetes 23:713-715. 17. Andres, R 1971. Aging and diabetes. Med Clin N Amer 55:835-846. 18. Andres, R, and J Tobin 1972. Aging, Carbohydrate Metabolism and Diabetes. Proc 9th Intl Cong Gerontology 1:276-280. 19. Kilo, C, N Vogler, and JR Williamson 1972. Muscle capillary basement membrane changes re- lated to aging and to diabetes mellitus. Diabetes 21:881-905. 20. 20. 22. The Problem 7 The Framingham Study Sept 1968. An Epidemiological Investigation of Cardiovascular Disease. Sections 10-22, US Government Printing Office, Washington, D.C. Keen, H, et al. 1965. Blood sugar and arterial disease. The Lancet 2:505-508. Ostrander, LD, et al. 1965. Relationship of cardiovascular disease to hyperglycemia. Ann Intern Med 62:1188-1198. - Be aS 3 Fe gs 3 y Nest 1% V padine : Chapter 2 Chapter 3 Chapter 4 Chapter 5 EPIDEMIOLOGY The Overall Problem and Its Impact on the Public 11 Harvey C. Knowles, Jr. Curtis L. Meinert Thaddeus E. Prout Economic Impact of Diabetes Paul S. Entmacher Improving the Organization of Care for the Chronically Ill Leona V. Miller Jack Goldstein John W. Runyan, Jr. The Computer in the Management of Diabetes Robert E. Bolinger 33 41 47 ge Sk 2! le a i Galen = ¥ Ww Jeng NY 2. at i Fe nie we ES gins TAEEER LS i phys oh . I : a Ball 18 4a E rf : eh wa 1h dell = ig enn fre 3 pi resi Fe Er SIE : a B I B = A na “ i ~ A] i t ¥ wld A : HRA 3 ny Sor a IL 1 m . a : oo 2 = Tp Goll Px TE Fae - n i vo DIABETES MELLITUS: THE OVERALL PROBLEM AND ITS IMPACT ON THE PUBLIC Harvey C. Knowles, Jr., Curtis L. Meinert, and Thaddeus E. Prout It is the purpose of this report to give estimates of the long-term problems of diabetes facing this country today. The estimates given will deal mainly with the epidemiology of dia- betes and hyperglycemia in the United States, time lost from activity because of diabetes, the risks of atherosclerosis and hypertension in the known diabetic population, the significance of hyperglycemia in the atherosclerotic population, the risks of small blood vessel disease in the known diabetic population, and the problems of neuropathy and of pregnancy in diabetes. Finally, the impact of society, needed areas of research and applied treatment, and the support necessary will be discussed. THE EPIDEMIOLOGY OF DIABETES Epidemiology of diabetes in the United States is unclear in many aspects. First, the preva- lence of diabetes is not known with certainty. Differences in testing procedures and criteria for diabetes have led to variations in reported results. Nevertheless, crude estimates of preva- lence can be made, and prevalences by measurements of blood sugar levels or by interview or medi- cal history in population samples are listed below. Cases per 1000 Source Year Age Known Newly Discovered Reference Oxford, Massachusetts 1946-47 15+ 11 13 (91) Tecumseh, Michigan 1959-60 20-79 18 (69) Nat'l Health Exam Survey 1960-62 18-79 18 (62) Sudbury, Massachusetts 1964 15+ 11 8 (71) Nat'l Health Interview Survey 1965-66 all 14.5 8 (28) Nat'l Health Interview Survey 1967 all 16.1 (64) Pima Indians ’ 1966 15+ 158 168 9) Despite variations in population sampling and in use of different criteria for diagnosis of diabetes, a common prevalence of approximately 2 percent emerges. Excluded from this estimate and included for contrast is the prevalence observed in the Pima Indians in Arizona. Just why the Pima population should have such a high frequency is unknown. The estimate of 2 percent in the non-Indian studies would yield a diabetic population of approximately 4.0 million in the United States if a total population of 200 million is assumed. Of the 4.0 million, it is likely that 2 to 3 million are now known, and one-half to one million have diabetes and remain to be - diagnosed. The prevalence of diabetes in children is apparently one in one thousand. About 389,000 new cases are reported yearly for the total population for all ages. It is estimated that there are about 5.5 million other persons in the United States destined to develop diabetes. Taking these figures into account, one can arrive at a total figure of approximately 10 million persons in the United States who have diabetes known, diabetes unknown, or will develop diabetes. These groups consist respectively of approximately 1.6, 0.6, and 2.8 percent of the population. Estimates by interview have been made of the ages of known diabetics in the United States. United States Public Health Services (USPHS) information obtained in 1965 and 1966 indicated 11 12 Diabetes Mellitus that at the time, 79 percent of male and 83 percent of known female diabetics were at least 45 \years old (28). Five percent were age 24 or less. Below age 45, the relative frequencies by sex appear equal, but after age 45, the rates for females are higher. Another poorly understood area of epidemiology concerns the age at diagnosis of diabetes (63). Eight percent of cases are diagnosed at age 24 or before, 22 percent from 25 through 44, 50 percent from 45 through 64, and 20 percent at age 65 or older. Only 4 percent reported their diagnosis to be made before age 15. Although it would seem that genetics play a role in the development of diabetes, the factors accounting for pancreatic failure at different times of life are anything but apparent. Observa- tions in other countries suggest that movement from rural to urban areas leads to an increased risk of diabetes. Increases in intake of fat and sucrose, as well as obesity, have been impli- cated as influential factors, but simply a high caloric intake may be a common denominator (87). Of particular interest at the minute, are the observations of a relation of development of juvenile-type diabetes with overt insulin failure to viral epidemics (54). The problem of genetic versus environmental influences on the development of diabetes and complications is well de- scribed in the studies of Pyke on diabetes and retinopathy in identical twins (82). In 1965, 33,000 deaths were listed as due to diabetes, and the disorder listed as the eighth leading cause of death (28). In 1972 and 1973 respectively, 39,000 and 36,000 deaths were so listed in the provisional statistics, and the disorder had advanced to the sixth cause (66). These figures may be conservative, however, for many deaths related directly or indirectly to diabetes may be attributed to other causes. Although a change in the coding classification system used for mortality statistics occurred between 1967 and 1968, from the Seventh to the Eighth Re- vision of the International Classification of Diseases, the increase cannot be attributed to this since the comparability ratio for diabetes mellitus between the two classifications was 0.9971 (67). The growth of the population and its changing age distribution have some effects, of course, on the diabetes mortality picture, but neither of these factors can account for the increase dur- ing this period. Both the crude and the age adjusted mortality rates have increased (66,67). In regard to survival with diabetes, the most extensive studies are those dealing with long- term follow-up of patients seen initially at the Joslin Clinic. Entmacher et al. in 1964 re- ported on 17,654 cases seen at the Clinic (31). Of great importance was the decline in deaths from ketoacidosis from 14.5 percent before 1930 to 1 percent in 1960. Vascular deaths rose in the same time from 46.8 to 76.6 percent. Tuberculosis deaths declined from 5.5 to 0.3 percent. Hirohata et al. calculated the cumulative survival rates in patients seen at the Joslin Clinic since 1939 and living in Massachusetts (42). The rates were expressed as percentages relative to those in Massachusetts life tables and are given according to age at first clinic visit and anniversary of follow-up. Age (years) Anniversary (years) Males Females 0-19 5 99.6 96.9 15 97.4 95.4 25 82.6 83.2 20-39 5 100.0 96.7 15 96.7 92.1 25 82.6 86.3 40-59 5 97.0 96.6 15 83.8 78.0 The Overall Problem 13 Age (years) Anniversary (years) Males Females 40-59. 25 75.0 55.1 60-79 5 98.4 88.4 15 77:1 62.8 25 62.9 35.1 The survival rates declined with duration of known diabetes. After age 40, they declined more rapidly, and those of women declined faster than those of men. A third follow-up study of cases seen at the same clinic was conducted by Kessler (47). SMR P All causes 1.93 <0.001 Cardiovascular disease 1.84 <0.001 Cerebrovascular disease 1.17 <0.05 Nephritis 1.27 >0.1 Tuberculosis 1.51 <0.05 Cancer 0.95 >0.1 Accidents 1.03 >0.6 Standardized mortality ratios (SMR), defined as the observed divided by the expected death rates, differed significantly for all deaths and for those due to coronary artery disease (CAD). Death rates from cerebrovascular disease and tuberculosis were probably higher. No increase in deaths from renal disease could be identified. It is to be remembered, however, that the causes are those given in death certificates. DISABILITY AND COST OF DIABETES The findings of the National Health Interview Survey of the year 1965-1966 indicated that 562,000 diabetics were disabled in some way from diabetes (28). There were 38.8 million days of restricted activity due to diabetes, 19.9 million of these days being spent in bed. The disabil- ity days due to diabetes per year per diabetic were as follows: - Restricted Activity Days Bed Days Other Total Men 6.7 5.2 11.9 Women 7.5 8.1 15.6 Eighty-two percent of the diabetics made visits to physicians in the year, and they averaged 6.6 visits each. In regard to hospital patients having diabetes, 1.7 percent of hospital dis- charges in 1971, exclusive of obstetrical and newborn discharges, had diabetes given as the first listed final diagnosis (65). This rate is equivalent to a discharge rate of 21.3 per 10,000 popu- lation. If a diabetes prevalence of 2 percent is assumed, then there are each year about 10 hospitalizations for diabetes per 100 diabetics in the population. Some of these persons may have more than one hospital episode in the year. In general, disability days due to all illnesses were 3 times more frequent among diabetics than those estimated for the United States population gener- ally (63). The same source provided also useful information for that time period (1965-1967) on the cost of diabetes to the American economy (28), although these figures need upward revisions in terms of present cost levels. The total was almost 2 billion dollars, with the division as follows: 14 Diabetes Mellitus Millions Estimated losses in earnings 1,344 Hospitalization 170 Nursing home care 45 Physician visits 90 Surgical fees 11 Insulin and oral drugs 148 Needles, syringes, etc. 8 Sugar tests (blood and urine) 51 Research 19 TOTAL 1,886 It is of interest that even with physician fees coming to 101 million dollars yeariy , about half a million diabetic patients, more than one-fifth of the total known diabetics, stated that they had not consulted a doctor during the course of the year. Those who did see a physician made, on the average, 6.6 visits during the year. LONG-TERM COMPLICATIONS Unfortunately, reliable information on rates of developing long-term complications of dia- betes are not at hand. Data on the natural history of disease have been collected from death certificates, hospital clinical and autopsy records, or obtained by surveys at clinic visits. These approaches provide crude estimates only, since they may deal with populations that happen to be available. Retrospective examination of records is an approach anything but satisfactory to obtain accurate facts. Prevalence surveys of living patients deal only with survivors. The use of prospective studies is a significant step forward in description of the course of diabetes, but so far there have been few efforts to measure events in diabetes precisely and express data in life table fashion (45). Even so, these few studies are limited to the populations which hap- " pened to be available when the studies were initiated. What ts urgently needed is a random sample of a diabetic population, as well as subjects with asymptomatic hyperglycemia, observed prospec- tively over a 20-year period with measurements of response variables made with sufficient preci- ston to allow for data interpretation. Atherosclerosis is the most common debilitating hazard faced by the diabetic patient. In- deed, coronary artery disease (CAD) is the most common cause of death in the middle-aged, stable- type diabetic. Bell, in his classical studies of diabetics at autopsy, listed cardiovascular disease as the major or contributing cause of death in 49 percent of 1,555 necropsies (7). Of the vascular deaths, CAD accounted for 18.5 percent, peripheral atherosclerosis tor 12.7 percent, and cerebrovascular disease for 7.7 percent. In a control group of 4,419 apparently nondiabetic autopsies, CAD was judged responsible for 24 percent of deaths. Of 5,009 deaths at the Joslin Clinic in the years 1960-1968, 77 percent were believed due to CAD (13). In contrast to this, vascular disease was held accountable for only 54 percent of all deaths in the United States in 1966 (93). Thus, in all studies, atherosclerosis stands out as a major health hazard in the natural history of diabetes mellitus, and CAD is the leading cause of death. Accordingly, CAD will be discussed in some depth. The Overall Problem 15 CORONARY ARTERY DISEASE Three recent reviews of coronary artery disease (CAD) in diabetes have appeared which contain references to most of the studies made in the past 30 years (14,33,86). The lesions differ little from those in the nondiabetic population and will not be discussed here. In the studies referred to, CAD prevalence varied from a low of 18 percent (7) to a high of 75 percent (81) in a study where a dye injection technique was used. On the average, the frequency of CAD at autopsy in the diabetics was 2.5 times that in the nondiabetics. Of less certain value are comparative data from surveys for CAD in living diabetic patients. Author Group Pathology Age No. Percent Liebow (56) 1955 | Diabetic CAD 10-90 383 42 MI 7 AP 10 Bryfogle (17) ..1957 Diabetic CAD >40 394 56 AP 12 Anderson (2) 1961 Diabetic CAD 23-88 100 55 Liebow (57) 1964 Diabetic CAD 40-70 58 33 UGDP (49) 1970 Diabetic CAD 20-79 1017 9.5 Kannel (44) 1961 Framingham CAD 30-62 4469 1.6 Epstein (32) 1965 Tecumseh CAD 16-70+ 5129 4.1 CAD - coronary artery disease MI - myocardial infarction AP - angina pectoris Prevalences at a given time have been reported as 42 and 33 percent by Liebow et al. (56, 57), 55 percent in a black population by Anderson et al. (2), and 56 percent by Bryfogle and Bradley in the Joslin Clinic (17). In the University Group Diabetes Program the frequency of CAD as evidenced by ECG change or history of angina within one year of diagnosis of diabetes in a middle-aged, stable-type of population was 9.5 percent (49). For comparison, the frequency of CAD in total populations in the Framingham study was 1.6 percent (44) and in the Tecumseh study 4.1 percent (32). It is of interest that recent observations in the Pima Indians, a race with an extraordinarily high prevalence of hyperglycemia (see above), the prevalence of CAD in the hyper- glycemics was only 5.3 percent in contrast to 3.3 percent in the normoglycemics (10). The preva- lence of angina pectoris in diabetics has been reported as 7 and 12 percent by Liebow et al. (56) and Bryfogle and Bradley (17), respectively. Certain items of interest come to light on reviewing the above studies. First, the sex dif- ferential of a higher risk of CAD in males in the general population is not apparent in the dia- betic samples. Diabetic women, prior to the menopause, have a prevalence equal to men of the same age. Indeed, there is some suggestion that after age 50 women might even be at greater risk (34). Second, CAD appeared to progress and cause symptoms at a younger age in the diabetic popu- lation . This premise is not yet certain, however, since sudden deaths from coronary occlusion in young men are not uncommon. Third, several items originally believed to add to the risk of CAD in the population at large do not seem to hold in the diabetic population. For example, in the majority of studies quoted above, there was insufficient evidence to substantiate obesity as a risk factor. In the studies of Pell and D'Alonzo there was a higher mortality rate in obese diabetics, however (75). Herein, the authors felt that the added risk was due to obesity and not a potentiating effect on diabetes. The role of elevated cholesterol as a risk factor for CAD in the diabetic was not evident in the Framingham study (35). Finally, the relation of CAD to 16 Diabetes Mellitus severity of diabetes as judged by insulin dose,:is uncertain. In the Framingham study, an increased mortality in diabetics taking insulin occurred only in the women. Indeed, a clear rela- tion between elevated blood sugar per se and the development of CAD has not yet been established. Three events do seem to increase the risk of CAD, however. First, all studies demonstrate a higher prevalence in patients with long duration of known diabetes. The known duration is a poor indicator of actual duration, however, since in many of the middle-aged, stable-type diabetic pa- tients prone to CAD, hyperglycemia may have existed without symptoms for many years prior to diag- nosis of diabetes. Second, as has been found in the whole population, all studies of hypertension and CAD in diabetes have reported.an increased prevalence of CAD when hypertension is present. Pell and D'Alonzo felt, however, that similar to the effect of obesity discussed above, the added risk was from hypertension per se and not from potentiation of the diabetic effect (75). Third, it has been proposed that there is overlap of diabetes and Type IV hyperlipidemia, that latter condition often being accompanied by atherosclerosis (11). The exact prevalences of elevated serum triglycerides and pre-beta band staining in the diabetic population are not clear at this time (80). Of particular importance in the course of myocardial infarction of the diabetic is the high immediate mortality and the low 5-year survival rate in those recovering from the initial attack. Immediate Five Year Total Mortality Survival Survival Author Year No. Percent No. Percent Percent Katz (46) 1949 57 51 Robinson (78) 1952 39 31 Cole (24) 1954 63 33 42 43 28 Bradley (12) 1956 83 58 40 40 17 Liebow {57) 1964 38 26 Partamian (72) 1965 205 38 127 38 24 Observations of hospitalized patients revealed immediate mortalities of 33 to 58 percent in initial attacks. In the control observations in four of these studies, the mortality was 20 to 28 percent, suggesting a better immediate survival in the nondiabetic. Of even more significance is the decrease in 5-year survival of those not dying in the first attack. In three studies, the survivals were only 38, 40, and 43 percent. When these death rates are combined with those of the initial attacks, the total 5-year survival rates are 28, 17, and 24 percent, respectively. In other words, a diabetic experiencing his first known myocardial infarction has roughly a 20 percent chance of being alive 5 years later. Data on the frequency of CAD in the younger insulin-dependent diabetic population unfortu- nately are meager and may be obscured by inclusion of other forms of myocardiopathy. Preliminary observations on the course of vascular disease in juvenile diabetes, with diagnosis made at age 16 or before, suggest that the risk of microangiopathy precedes that of macroangiopathy, and that those at risk may die of renal failure before the clinical appearance of CAD (51). After the 30th year, the risk of small blood vessel disease decreases, but at this time the risk of athero- sclerosis increases. Whether this new risk is related more closely to age or to duration of known diabetes cannot be determined at this time. In summary, CAD in diabetes is characterized by a high prevalence, an equal risk by sex, occurrence at an earlier age, and a high death rate at the time of attack with a low 5-year The Overall Problem 17 survival rate. It occurs mainly in the middle-aged, stable-type of diabetic, and the high prevalence makes it the most frequent and hazardous risk in the diabetic population. CAD is unusual in juvenile-type diabetes, but may become more prevalent as longevity increases in the juvenile diabetic population. PERIPHERAL VASCULAR DISEASE Peripheral vascular disease, characterized by atherosclerosis and obliteration of the arte- ries in the legs, is very common in the diabetic population. Monckeberg's sclerosis, calcifi- cation of the vessel media, is also common, but is asymptomatic because the vessel lumen remains patent. Warren observed a frequency of 34 percent of peripheral arterial disease at autopsy in dia- betics at autopsy (86). Bell found approximately 13 percent of the diabetic population had periph- eral gangrene at death (7). In clinical studies, prevalences of 58 and 59 percent have been ob- served (16,53). In the Framingham study the morbidity of intermittent claudication in the dia- betics was 4.5 times that expected (35). Peripheral atherosclerosis occurs at a younger age in the diabetic population than in the general population. Similar to the observations on CAD, there does not appear to be a relation to severity .of diabetes, though one would say it was more common with extended duration. Nor can a relation to severity of insulin failure as judged by insulin dosage be established. Studies have indicated that following amputation of one leg, the other leg is involved in 50 percent of cases in 2 years, and in 95 percent in 5 years (36). Moreover, only 36 percent of amputees may be alive in 5 years (90). Juvenile type diabetics with diagnosis before age 16 seem rarely to develop arterial occlusion below the knee. Monckeberg's sclerosis is very common in the juvenile population, however, and White found an 83 percent prevalence of medial calcifi- cation in long-standing juvenile diabetes (88). CEREBROVASCULAR DISEASE Whether or not there is increased frequency of cerebral atherosclerosis in diabetes is un- settled. Warren and LeCompte have suggested that it may be uncommon simply because diabetics often die with CAD before cerebral sclerosis can develop (86). Grunnet found two-thirds of 107 diabetics to have at autopsy evidence of cerebral sclerosis, but control data were difficult to acquire (38). Bradley states that 12 percent of diabetics may die from cerebrovascular accidents, a prevalence not different from that in the total population (13). Bell ascribed 8 percent of diabetic deaths to be from cerebrovascular disease as ascertained at autopsy (7). On the other hand, the morbidity of cerebrovascular disease in the diabetics in the Framingham study exceeded the expected by a factor of 2.4 (35). It would seem unlikely that the different prevalence of atherosclerosis in the diabetic and nondiabetic population would exclude the cerebral arteries, and that the concept of early death from other diabetic complications could explain the lack of overt evidence for an increased frequency. HYPERTENSION In this discussion, hypertension is referred to as that known as ''essential' and is apart from the hypertension associated with renal disease. There have been numerous studies to determine if hypertension is more frequent in the dia- betic population. The usual criticism of poor standardization of measurement, unsatisfactory 18 Diabetes Mellitus sampling of the diabetic population, and lack of suitable control data can be applied in all. Nevertheless, there is the clinical impression that hypertension is more common than exptected and that women and the elderly may, in particular, be at risk for elevated blood pressure. A recent study of DuPont employees indeed supports the view of an increased prevalence in the middle-aged stable-type diabetic population (74). Listed below are prevalences of hypertension observed in various diabetic groups. Author Year Patients No. Percent White (88) 1956 Juvenile diabetes, 879 53 15+ years Bryfogle (17) 1957 Clinic, all ages 394 36 Pell (74) 1967 Industrial employees 662 37 Liebow (56) 1955 Clinic, all ages 383 43 Anderson (2) 1961 Clinic, all ages, 100 59 (black) Brandman (16) 1953 Clinic, all ages 264 24 UGDP (49) 1970 Clinic, stable type 1017 31 diabetes at diagnosis All the studies deal for the most part with the general diabetic population except for that of White, which concerns only long-standing juvenile diabetes (88). The frequencies given range from 24 to 59 percent, the differences being due to variations in age at time of measurement and in acceptable blood pressure levels for hypertension. In most instances, women appeared to be at a greater risk than men. Accelerated or malignant hypertension has not been reported as a common event in diabetes. Control data on nondiabetic patients was unavailable or inadequate except for the study by Pell and D'Alonso (74). HYPERGLYCEMIA IN CORONARY ARTERY DISEASE In recent years attention has been directed to the prevalence of hyperglycemia in populations with vascular disease. By hyperglycemia is meant elevated blood sugar, but not necessarily to the level of diabetes. Unfortunately, the definition of hyperglycemia varies from author to author. In 47 patients with various forms of severe atherosclerosis, Waddell and Field found the OGTT to be 'mormal" in only 15 percent (84). Bartels and Rullo observed abnormal glucose toler- ance in 77 percent of a group of patients with peripheral vascular disease and no known diabetes (5). Most investigations of blood sugar levels in atherosclerosis have been made in CAD, how- ever. Nine studies of oral glucose loading, particularly in the recovery phase of myocardial in- farction yielded hyperglycemia ranging from four to 77 percent in frequency, Again, many of the studies were hindered by lack of satisfactory control data, and often they were carried out at varying periods after a myocardial infarction. The significance of hyperglycemia in the atherosclerotic population remains to be determined. There is no question of the increasing risk of atherosclerosis in known diabetes where hyper- glycemia has been established previously. The reversed situation of mild hyperglycemia in the atherosclerotic population allows for speculation on the relation of hyperglycemia and athero- sclerosis. Knowledge of whether or not a cause and effect relation occurs will be of the great- est importance in the direction of efforts to curtail vascular disease in diabetes. RETINOPATHY AND BLINDNESS Disorders of the eyes with loss of visual acuity are a common and serious complication of The Overall Problem 19 diabetes. Recent reviews are those of Bradley (15), Leopold (55), and Caird (21). ZL In 1935 Waite and Beetham reported their findings of eye abnormalities in 2,002 diabetics and observed retinal hemorrhages in 18 percent and exudates in 10 percent (85). Since then, many prevalence studies of retinopathy in clinic populations have appeared, with frequencies ranging widely because of differences in ages and durations of diabetes in the patients. In juvenile diabetes with disease known of 10 or more years, frequencies of 19 to 100 percent have been re- ported and are listed in the review of Knowles et al. (50). Those with the lowest rates, 19 per- cent (25) and 23 percent (43), are of interest because they included patients with diabetes diag- nosed before age 10. At the other extreme, White found lesions in 90 percent of juvenile diabetic patients with 30 or more years of known duration (88). All these surveys are hardly comparable, however, because of the variations in patient ages and especially in durations of known diabetes. Burditt et al. has attempted to assemble published prevalence data in relation to age at diagnosis and duration of known diabetes (18). In general, the frequency of retinopathy increases with duration regardless of age at diagnosis, with frequencies ranging from 42 to 82 percent at 24 years duration. In one long-term prospective study of juvenile diabetes, a somewhat lesser fre- quency of 62 percent at 25 years duration was found (51). This is a lower figure than that given by White (88) and may be related to the population studied. The former data are from a complete population observed from the time of diagnosis of diabetes, whereas many of the other studies are of referred populations which may be biased by reasons of referral. Finally, it is of interest that in the juvenile diabetics, the risk of retinopathy appears to decrease after 30 years (51). The prevalences of retinopathy given above include all degrees of retinopathy from simple red dots to neovascularization with gross hemorrhage and scarring. Neovascularization, or malig- nant retinopathy, is more common in the long-term younger insulin dependent diabetic. In a study of a group of 847 patients with malignant retinopathy, Root et al. found over 80 percent of cases to be in the age range of 20 to 59 (79). White, in her studies of diabetes with onset in child- hood, found a frequency of 53 percent at ages 30-39 with an increase to only 58 percent in the 40 year age group (88). Visual impairment usually occurs in a few years after the appearance of new blood vessel formation. The time for progression to blindness is quite variable, some cases progress rapidly in a year, some remain stationary, and others, estimated to be 10 percent in number by Beetham (6) and Davis (27), show regression of lesions. In Beetham's series of 351 cases observed over 4 years, 30 percent had become legally blind, 7 percent totally blind, and 10 percent had im- proved (6). Retinopathy in the older person appears to progress to gross visual impairment at a more rapid rate than. in the younger person, as exemplified in the data of Caird et al. (20). The impact of diabetic retinopathy on public health lies in the number of patients in the total population developing visual impairment from diabetes. The prevalences of blindness for all causes in the United States have not been clearly established because the reporting of blind- ness is not compulsory. Nor is there certainty as to the exact cause of blindness in many in- dividuals. Blindness in diabetics is estimated to be 10 to 28 times as frequent as in the gen- eral population (21). One source of valuable statistics on the cause of blindness is that of the Model Reporting Area for Blindness Statistics which indicates that diabetic retinopathy is among the top four causes of blindness (61). Diabetic retinopathy accounted for 12 percent of new cases of blindness added to this register in 1966. The other three common causes of blindness 20 Diabetes Mellitus were maculopathies, lenticular problems and glaucoma; each of these may have diabetes as a major contributor. Blindness from diabetic retinopathy occurred at a younger age than blindness from the other three causes. The average age at the start of blindness due to diabetic retinopathy in the total diabetic population was 60 years compared with an average age of 78 for patients with macular degeneration, 72 years for patients with cataract, and 73 years for patients with glaucoma (61). Diabetic retinopathy is reported to be the most common cause of newly reported blindness in the age group 41-60 years and the second most common cause in the age group 21-40 years, second only to congenital defects (40). The National Society for the Prevention of Blindness estimates that 44,660 individuals were blind from diabetes in the United States in 1962 and that 4,480 persons became blind from diabetes during that year (40). Thus, when blindness is viewed in all of its many aspects, it is clear that: a. Diabetes is the major systemic illness causing blindness. b. Diabetes is the major contributing factor to blindness for reasons other than that due to diabetic retinopathy alone. c. Diabetic patients have from 10 to 28 times the amount of blindness found in the normal population. d. Diabetes is the leading acquired cause of blindness in individuals in the most produc- tive years of their lives. Although diabetic retinopathy rarely appears early in the course of juvenile diabetes, usu- ally being evident at 15 to 20 years after the diagnosis of diabetes, it becomes an almost in- evitable consequence of diabetes with onset at an early age. Caird has estimated that chances of visual impairment or blindness sufficient to give rise to difficulty with employment rises with age from 3 percent in those under 30 at diagnosis of diabetes to 32 percent in those over 60. One can thus appreciate the economic cost to society (21). The pathologic changes in the eye often are associated with changes elsewhere in the body. For example, relationships between patient survival and severity of retinopathy have been studied by Thomasen (83). In groups of patients matched for age and sex, the observed mortalities for the nondiabetic populations and the diabetics with microaneurysms only were 86 percent and 90 percent, respectively, in a 7-year period. On the other hand, percent survival in diabetic patients with more severe retinopathy during the same period was 41 percent for hemorrhages and/or exudates and 32 percent for malignant retinopathy, respectively. Certified causes of death in 660 cases of proliferative diabetic retinopathy studied at the Joslin Clinic were, in order, nephropathy, coronary occlusion, arteriosclerosis, and cerebral vascular accident (15). The relationship between the severity of retinopathy and glomerular disease in renal biopsies is well known. The triopathy of retinopathy, neuropathy, and nephropathy occurring in the same patient and increasing in percentage with the duration of diabetes is also well established. The cluster of these pathologic events is of great importance, for it emphasizes the systemic nature of the pathology of diabetes, the relentless progression of this disorder in young patients, and the increasing frequency with which other major disabilities such as neuro- pathic ulcerations of the feet, chronic renal failure, or coronary heart disease may accompany significant retinopathy. Many of the diabetic blind have lost their vision in the third and fourth decade of life, a time when they are most productive. There is loss of income with The Overall Problem 21 personal hardship that the diabetic family, as well as the diabetic patient, suffers. Not of the least in importance is the cost of services to the blind that society bears. CATARACTS In their early extensive studies, Waite and Beetham found a prevalence of cataracts of 50 percent in 1,732 diabetic patients (85). The prevalence in a control group of 526 apparently nondiabetic patients was 57 percent. McGuinness also found no increase in diabetics (59), but Caird is of a contrary opinion (21). It is generally accepted that cataracts are unduly common in the juvenile diabetic population, and Knowles et al. list reported frequencies ranging from 2 to 47 percent depending on ophthalmoscopic or slit lamp examination (50). Cataracts in the juve- nile population are often of the metabolic or '"snowflake' type, while those in older populations usually are the ''senile'" type and similar to those in the nondiabetic. The frequency in women appears to be somewhat higher than that in men between ages 40 and 70. OCULAR PRESSURE It is not certain if primary glaucoma exists in the diabetic population in increased fre- quency (15). In observations of primary open-angle glaucoma, Armstrong et al. found, however, a frequency of 4.8 percent in 393 adult diabetics in contrast to 1.8 percent in 280 controls (3). Intraocular tension is reported to be higher in the diabetic population, especially in juvenile diabetes (15). The occurrence of secondary glaucoma following proliferative retinopathy is well known and is found in about 10 percent of those with vitreous scarring. RENAL DISEASE Renal disease in diabetics is a serious hazard to patient activity and longevity, particu- larly in the young diabetic. Recent extensive reviews have been published by Balodimos (4) and Rifkin (76). The most common lesions are diabetic glomerulosclerosis and arteriolonephroscler- osis. Pyelonephritis occurs also and is occasionally accompanied by medullary necrosis. An in- creased frequency of asymptomatic urinary tract infection has been claimed but conflicting evi- dence has not supported this. The lesions incapacitating the diabetic subject are glomerul- osclerosis and occasionally pyelonephritis. Other common forms of chronic renal disease do not appear to occur in diabetes beyond that expected in the general population. Glomerulosclerosis may exist to some extent in all diabetics. At the moment the diagnosis is made by clinical course and by light microscopy at renal biopsy or autopsy. It is not always possible to delineate early mesangial lesions. At autopsy prevalences ranging from 15 to 82 per- cent have been reported with the mean prevalence being about 30 percent (4). Bell, in his study of 1,465 autopsied diabetics found 19.5 and 30 percent frequencies of glomerulosclerosis in men and women respectively (8). In each sex, two-thirds of the cases were of the diffuse type only, and one-third included nodules as well. Kimmelstiel reported an autopsy frequency of 17 percent of nodular glomerulosclerosis (48). Heptinstall reported either diffuse or nodular lesion oc- curred at autopsy in 50 percent (41). In general, there have been more female than male cases. Warren and LeCompte state that renal failure caused death in 12 percent of diabetic cases at the New England Deaconess Hospital (86). In a study of the causes of death in 6,800 patients seen at the Joslin Clinic between 1956 and 1964, 389, or 6 percent, died of nephropathy (4). There were 208 men and 181 women, and almost half of the cases were of juvenile type with diag- nosis made before age 20. Indeed, life insurance data indicate that death from renal disease is 22 Diabetes Mellitus 17 times more frequent in the diabetic population than in the nondiabetic population (31). In living diabetics, Bryfogle and Bradley estimated the prevalence of glomerulosclerosis to be 10 percent in a clinic population of diabetic patients of all ages (17). The younger dia- betic population is very much at risk for glomerulosclerosis, however, and the prevalences in those with known diabetes of 10 or more years duration range from 4 to 100 percent (50). In her classical studies of long-term juvenile diabetes, White quotes the frequencies of glomeruloscler- osis as evidenced by proteinuria to range from 18 percent at 15 to 19 years' duration to 63 per- cent at 35 to 39 years' duration (88). In a recent report of prospective observations of 167 juvenile diabetics followed beyond 10 years, the cumulative risks of proteinuria at 20, 25, and 30 years' known duration were 26, 37, and 48 percent, respectively (51). These figures are less than those given by White and could be explained again by the populations studied. Data on survival with proteinuria have been developed by Caird (19). He found that at 5 years, the survival rates were of diabetics with proteinuria, 65 percent; diabetics without pro- teinuria, 73 percent; and the control population, 83 percent. Pell and D'Alonzo also published data on the relation of proteinuria and survival (75). In diabetics in an industrial population, they found that the 10-year death rates of diabetics with and without proteinuria were 39 and 23 percent, respectively, in comparison to a control rate of 10 percent. Finally, it should be noted that in juvenile diabetics, proteinuria is very unusual before 10 years' duration of known diabetes, and similar to the risk of retinopathy, is unlikely to appear de novo after 30 years of known duration (51). In line with this, White observed in 73 juvenile diabetics of 40 years' duration that though 18 had evidence of nephropathy, 12 had pyelonephritis (4). When azotemia appears in the juvenile diabetic, death follows usually within 3 years (52). Urinary tract infection poses another renal problem for the diabetic patient. Studies of asymptomatic urinary infection in the diabetic population have been made with conflicting results (4). Studies of pyelonephritis made at autopsy suggest a higher rate in diabetes, with preval- ences of 7 (78), 12 (29), and 36 percent (86) being reported in contrast to control rates of 2 and 3 percent. Variation in the criteria used for histologic diagnosis of chronic pyelonephritis likely accounts for the different prevalences. The relation between urinary infection and pyelonephritis is not clear. For example, in a study of renal biopsies in 80 advanced insulin-dependent diabetics undergoing pituitary ablation for progressive retinopathy, 7 percent were found to have chronic pyelonephritis and 12 percent had positive urine culture (39). No relation was found, however, between urine culture, tissue culture, and biopsy appearance. Finally, it is estimated that medullary necrosis, a potentially lethal renal lesion, is present in 4 to 9 percent of autopsies of diabetics (1). NEUROPATHY Diabetic neuropathy implies abnormal function of peripheral nerve pathways in the diabetic patient. Almost any pathway, somatic or autonomic, motor or sensory, can be involved. The dis- turbed functions can be ones of increase or decrease, and the course can be fluctuant or steady and end in permanent dysfunction or remission. The relation of neuropathy to other aspects of the diabetic syndrome is not understood and will not be discussed here. Recent reviews of neuropathy have been given by Colby (23) and Ellenberg (30). The vari- ability of its course makes it almost impossible to determine its frequency, and it is likely The Overall Problem 23 that most diabetics at some time or other experience some degree of abnormality. Indeed, with very precise nerve conduction measurements, it might be possible to disclose dysfunction in all diabetics. In earlier reviews, prevalences based on gross evaluation varied from 0 to 93 per- cent (37). Danowski et al. examined 374 consecutive clinic patients and failed to find an ankle reflex in 30 percent, decreased vibratory sense in 44 percent, and painful neuropathy in 9 per- cent (26). The difficulties in judging the excess of the milder neuropathic symptoms and signs in a diabetic population is apparent in the study of Mayne in England (58). He evaluated 220 diabetics and 110 control subjects of comparable age and sex ratio and found similar symptoms and signs in the controls, though of significantly less frequency. The loss of productivity time in the diabetic because of neuropathy cannot be determined. In the long-term juvenile diabetic population, about 10 percent will be incapacitated at a given time with pain, trophic ulcer, ataxia, or diarrhea. The loss of position sense makes it all the more difficult for the patient with failing vision to ambulate. Orthopedic corrective measures have been developed for trophic foot ulcers, but little else is available to alleviate the other neuropathic syndromes. PREGNANCY It is estimated that diabetes is listed as a hospital discharge diagnosis with pregnancy in about 1 in 300 discharged pregnant patients (89). This figure is conservative because many pa- tients with unknown gestational or transient diabetes may be overlooked. At the Cincinnati General Hospital in 1972, there were 66 admissions of patients with diabetes among 3,082 preg- nancy admissions, a frequency of 2.1 percent. In 1971 the frequency was 1.7 percent. These figures fall probably below the actual rates in view of the failure to diagnose gestational dia- betes. For present purposes, diabetes in pregnancy will be looked on as either known prior to pregnancy, or developing during pregnancy with remission after completion of pregnancy. This latter type has been called gestational diabetes (70). Wilkerson and O'Sullivan conducted oral glucose tolerance tests in 752 unselected pregnant women without known diabetes (92). Using criteria for the diagnosis in the population at large, the authors noted prevalences of ab- normal tests of 4.8 percent in the second trimester and 7.3 percent in the third trimester. In a study of glucose tolerance after delivery in gestational diabetes, O'Sullivan found 7 percent of gestational diabetics to remain diabetic after completion of pregnancy, and 66.7 per- cent to develop established diabetes in 5 and a half years (70). The 7 percent remaining dia- betics could have had diabetes established prior to pregnancy and therefore would not be classi- fied according to the strict criteria of gestational diabetes, where it is assumed diabetes did not exist before pregnancy. Pregnancy can influence carbohydrate metabolism in ways which will not be apparent in the nondiabetic person, but which will affect glucose tolerance in the patient with prediabetes or established diabetes. There is antagonism to insulin with failure of compensation by the pan- creas of prediabetic women and resulting rise in insulin requirement in established diabetic women. There is little evidence for residual effects in the diabetic mother, however. It is not believed that diabetic glomerulosclerosis is worsened, though background retinopathy can increase and decrease in pregnancy, and rapidly progressing neovascularization has rarely been 24 Diabetes Mellitus reported to occur in pregnancy. The natural history of diabetic retinopathy in pregnancy has not yet been fully explored, however. In regard to the effects of diabetes in pregnancy, there is no evidence for infertility nor for an increase in fetal loss in the first trimester. It is possible, however, that the frequen- cy of second trimester spontaneous abortion is greater. Most of the problems take place in the third trimester, and are exemplified in a study of 705 diabetic pregnancies made in Cleveland in 1964 (60). The percent abnormalities observed were compared to the nondiabetic pregnant popula- tion. Nondiabetic Percent Diabetic Percent Toxemia 4.4 20 Hydramnios 0.3 20.4 Lethal defects 0.5 2.0 4,000 gm or >infant 0.7 37 Perinatal death 18.3 In regard to the effects of diabetes on the fetus, there are increased risks of excessive weight from fat accumulation, congenital defects and stillbirth, and on the delivered infant of prematurity, serious hypoglycemia, hyperbilirubinemia, and respiratory distress syndrome. These conditions are discussed elsewhere. Good management of diabetes during pregnancy, coupled with selection of optimal time of de- livery by inducement or section, have done much to give a favorable outcome. The survival rate of the viable infant will vary according to the stage of diabetes. Data collected on outcome of 870 diabetic pregnancies treated at the Joslin Clinic give an overall fetal survival of 90 per- cent (89). The studies of Oakley (68) and Pedersen (73) in particular, emphasize the value of diabetic control. In essence, the impact of pregnancy on the diabetic woman is significant in that there is pregnancy, fetal and infant morbidity beyond the expected, though there is little damaging effect on the mother. Considerable time and effort are required by several physicians, as well as the patient, to assure the best outcome possible. Improved medical, obstetrical, and neonatal manage- ment have done much to increase fetal and infant survival. Two problems which so far defy cor- rection are the spontaneous delivery initiated too early for infant survival and stillbirth. The. first can be attacked through further advances in premature care. The second will require re- search into the causation. DELIBERATIONS The high prevalence, inconveniences of treatment, risk of debilitating consequences, and de- creased longevity qualify diabetes mellitus as a public health problem of high priority. Insulin became available in 1922, and the event was hailed as the solution to the problems of diabetes. Not appreciated at the time, however, was that interference in normal activity could be produced by treatment and short-term control complications, nor that physical failure and shortened life span could result from long-term complications. The public looked on insulin as the panacea, and the mistaken concept remains today that the diabetic is handicapped only by his treatment annoy- ance and occasional episodes of uncontrolled diabetes. It has been the purpose, so far, to de- scribe the status of the diabetic patient in the last decade. Emphasis from the epidemologic standpoint has been placed on the deficiencies in prevention and treatment in all areas of The Overall Problem 25 diabetes. It is hoped that efforts can be stepped up to fill gaps in knowledge in cause and prevention, and in corrective treatment where cause is unknown and prevention not feasible. To begin with, the estimate of 2 percent prevalence of diabetes is probably near the truth. Nevertheless, there are no prospective studies of glucose tolerance in selected populations other than that of the USPHS studies initiated in 1947 in Oxford, Massachusetts. Unfortunately, at the time the methods were crude. A similar study with standardized glucose tolerance testing and long-term follow-up would be of help to predict the future diabetic. In fact, of great value at the minute would be agreement on a single testing approach with criteria for the diag- nosis of diabetes, well knowing that the criteria may be changed as information accumulates. Also, the development of standardized screening criteria for use in everyday medical practice to determine diabetes or need for glucose tolerance testing would be most helpful. Studies to give data in these areas could be initiated by national agencies and appropriate departments of schools of medicine and public health. The ultimate value would be to mark the unknown diabetic before he seeks medical attention for an event which might have been preventable, such as ketosis and foot infections. Finally, such studies would be helpful in defining the sex ratio in dia- betes, knowledge of which would be helpful in investigations of the causes of different aspects of diabetes. The reasons for the development of diabetes are anything but understood. As mentioned be- fore, the aggregation of diabetes in families has led to a general view that a genetic influence must be present for diabetes to develop. The studies of Pyke of identical twins with strong family history of diabetes becoming concordant for diabetes in a short time and developing reti- nopathy in contrast to twins with less family history and less concordance for diabetes and reti- nopathy, support the concept of both genetic and environmental influences (82). Observations of the development of glucose intolerance made prospectively with notation of life events in fami- lies with strong diabetic history could yield information related to the course of insulin fail- ure. Clarification of the mechanism of inheritance plus determination of influential environmen- tal factors would be the first step towards development of a prevention program in diabetes. The number of days of normal activity lost from uncontrolled diabetes are small in the in- sulin independent population, but large in those taking insulin. Theoretically, significant ke- tosis and hypoglycemia should never occur. Such a goal is unattainable, but increase in physi- cian and patient education could do much to prevent wasted time. Studies of the reasons for time lost and effects on family, occupation, and environment could disclose areas where more intensive education is needed. Applied research to improve education could benefit markedly the economics of the young diabetic with fluctuant diabetes. It is likely that the 20 days per year of hospi- talization per 200 diabetic patients could be lessened. It is in the area of vascular disease that the diabetic develops most often his debilitating complications, however. Although the clinical vascular syndromes are well described, their inci- dence rates and natural history are not. To clearly define these, one would need a population of diabetics to follow prospectively for 25 years. Without parameters of the course of events, it is very difficult to evaluate treatment regimens of questionable effects, and it is unlikely that a treatment not requiring a clinical trial will be available in the near future. The only avail- able studies of prospective data are one in juvenile diabetes (51) and one in stable type (49), both of which have too few patients to determine other than gross incidences and trends. It is 26 Diabetes Mellitus unfortunate that large diabetic centers in this country have not used prospective approaches starting 20-some years ago so that natural history data now would be becoming available hope- fully for future intervention studies. Large gaps of knowledge exist concerning diabetic atherosclerosis. As mentioned before, cardiovascular disease is the leading serious hazard facing the diabetic. Many population studies of hyperlipidemia and cardiovascular disease are being conducted in lipid centers in the United States, but they are not directed toward diabetes as a subject. Peripheral atherosclerosis markedly affects the older diabetic. It is possible that diabetic microangiopathy is present, and detailed histologic studies in extremities are needed, similar to those carried out in the eye, kidney, and muscle. Similarly, neuropathologic observations are needed to determine if cere- bral atherosclerosis is increased, and if other vascular pathology is present. Of particular in- terest is the high prevalence of hypertension in diabetes apart from that of renal origin. In- vestigations of the pathogenesis are needed to determine if there is difference from essential hypertension in the nondiabetic population. The relation of microangiopathy to the course of diabetes remains one of the greatest of un- solved problems. Diabetic retinopathy is a leading cause of blindness before age 50, and renal failure is the most common cause of death in the juvenile diabetic. Yet the cause of microangiop- athy is unknown. Indeed, it is not known if mechanisms are the same in the development of reti- nopathy, glomerulosclerosis, and muscle capillary basement thickening. More extensive studies on the metabolism of diabetic vessels are urgently needed to disclose the factors causing the struc- tural alterations. The natural history of retinal neovascularization is far from clear, and it would be extremely helpful to disclose factors affecting its course as seen in spontaneous regres- sions, pregnancy, and perhaps after pituitary ablation. At the moment, there is no proven treat- ment of value for retinopathy. Pituitary ablation has been largely abandoned, and photocoagula- tion is still in a trial stage. Other approaches to the problem of retinopathy must be developed. Of greater importance is the renal lesion of diabetes, for this is the most serious hazard for the young diabetic. Currently there is no conservative approach to its treatment, but preliminary ex- perience with transplantation suggests that it may be successful in diabetes. Support for this should become available for Centers with large juvenile diabetic populations. The pathogenesis of neuropathy will require intensive investigation. Current studies of the metabolism of neuronal tissue must be extended to find where the pathologic process fits in the . general chemical derangements of diabetes. Orthopedic and physical medical and rehabilitative management can be advanced for the neurotrophic foot ulcer. The neurogenic bladder must be re- lieved, and corrective if not preventive approaches be developed for diarrhea, ataxia, and the peripheral somatic neuronal syndromes. The complications of pregnancy and the newborn of the diabetic mother have been lessened con- siderably by more exact treatment of diabetes, accurate time of delivery, and in particular, by marked increase in the skills of newborn care. Nevertheless, 5 to 10 percent of cases result in perinatal loss. Research into the mechanisms of premature delivery, stillbirth, and the respira- tory distress syndromes are desperately needed. In addition, a more intensive search should be made for the gestational diabetic, for she well may represent the commonly occurring female- middle-aged, stable-type diabetic. There has been extensive controversy as to whether the complications result from insulin The Overall Problem 27 failure and its consequences, whether they develop apart from inadequate insulin effect, or whether there are combinations thereon. Many. observers believe from clinical observation that patients with more complete normalization of the blood sugar have fewer complications. In- duced diabetes in animals has produced small blood vessel lesions resembling human diabetic microangiopathy. Furthermore, enzyme activity accounting for thickened glomerular basement membrane in alloxan diabetic rats has been defined. On the other hand, no satisfactory clinical trials demonstrating that control of diabetes diminishes the risk of vascular complications have been made. Also, it is not certain if the experimental vascular lesions in animals are the equivalent of those in man. Finally, microangiopathy has been found without demonstrable insulin failure at the time. Many treatment approaches with dietary control and insulin adjustment have been developed in the hope of ameliorating and even correcting some of the long-term complications in diabetes. But nothing has so far been successfully applied in this country. Even if it is established that insulin failure and uncontrolled diabetes lies at the base of complications, it is highly unlikely that satisfactory correction of the intermediary derangements can be brought about with present-day methods. Current studies at the animal level indicate that diabetes can be relieved in the experi- mentally induced diabetic model. Investigations on isolation, preservation, and proper site of implantation of islets in primates would be the next logical step, followed by observations on the course of the recipient in regard to carcinogenic and rejection risks. If these studies would be accomplished, a prospective clinical trial could be mounted in man to determine the effects on the course of vascular disease. Diabetes mellitus is one of the most complex of medical states. Insulin deficiency is the major known abnormality, but other hormonal impairment undefined as yet may be concerned. In- deed, genetic input may be more influential than realized. Accordingly, it is difficult to judge just what areas would be most profitable to pursue at this stage, since many are hindered by limitations in methodology. Nevertheless, investigations in the mechanisms of insulin failure, the application of implanted pancreases, and clarification of altered tissue metabolism would seem profitable. The impact on society of diabetes and its disease consequences is sufficiently great to warrant a marked increase in efforts to alleviate its public health problem. REFERENCES 1. Abdulhayoglu, S, and A Marble 1964. Necrotizing renal Papillitis (Papillary Necrosis) in diabetes mellitus. Am J Med Sci 248:623-632. 2. Anderson, RS, A Ellington, and LM Gunter 1961. The incidence of arteriosclerotic heart v" disease in negro diabetic patients. Diabetes 10:114-118. 3. Armstrong, JR, RK Daily, HL Dobson, and LJ Girard 1960. The incidence of glaucoma in diabetes mellitus. A comparison with the incidence of glaucoma in the general population. Am J Ophthalmology 50:55-63. 4. Balodimos, MC 1971. Diabetic Nephropathy. In Joslin's Diabetes Mellitus. Editors, A Marble, P White, RF Bradley, and LP Krall. 11th edition, Philadelphia, Lea § Febiger, pp 526-561. 28 Diabetes Mellitus 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. Bartels, CC, and FR Rullo 1958. Unsuspected diabetes mellitus in peripheral vascular disease. New Eng J Med 259:633-635. Beetham, WP 1963. Visual prognosis of proliferating diabetic retinopathy. Brit J Ophthalmology 47:611-619. Bell, ET 1952. A postmortem study of vascular disease in diabetics. Arch Path 53:444- 455. Bell, ET 1953. Renal vascular disease in diabetes mellitus. Diabetes 2:376-389. Bennett, PH, TA Burch, and M Miller 1971. Diabetes mellitus in American (Pima) Indians. Lancet 2:125-128. Bennett, PH 1973. Personal communication. Bierman, EL, and D Porte, Jr. 1968. Carbohydrate intolerance and lipemia. Ann Int Med 68:926-933. Bradley, RF, and JW Bryfogle 1956. Survival of diabetic patients after myocardial in- farction. Am J Med 20:207-216. Bradley, RF 1971. Coronary Artery Disease. In Diabetes Mellitus: Diagnosis and Treatment. Editors, SS Fajans, and KE Sussman. Volume III, New York, American Diabetes Association, pp 295-303. : Bradley, RF 1971. Cardiovascular Disease. In Joslin's Diabetes Mellitus. Editors, A Marble, P White, RF Bradley, and LP Krall. 11th edition, Philadelphia, Lea § Febiger, pp 417-477. Bradley, RF, and E Ramos 1971. The Eye and Diabetes. In Joslin's Diabetes Mellitus. Editors, A Marble, P White, RF Bradley, and LP Krall. 11th edition, Philadelphia, Lea §& Febiger, pp 478-525. Brandman, O, and W Redisch 1953. Incidence of peripheral vascular changes in diabetes mellitus: A survey of 264 cases. Diabetes 2:194-198. Bryfogle, JW, and RF Bradley 1957. The vascular complications of diabetes mellitus. A clinical study. Diabetes 6:159-167. Burditt, AF, FI Caird, and GJ Draper 1968. The natural history of diabetic retinopathy. Quarterly J Med 37:303-317. Caird, FI 1961. Survival of diabetics with proteinuria. Diabetes 10:178-181. Caird, FI, AF Burditt, and GJ Draper 1968. Diabetic retinopathy. A further study of prognosis for vision. Diabetes 17:121-123. Caird, FI, A Pirie, and TG Ramsell 1968. Diabetes and the Eye. Oxford and Edinburgh, Blockwell Scientific Publication. Clawson, BJ, and ET Bell 1949. Incidence of fatal coronary disease in nondiabetic and in diabetic persons. Arch Path 48:105-106. Colby, AO 1965. Neurologic disorders of diabetes mellitus. Part I. Diabetes 14:516-525. Cole, DR, EB Singian, and LN Katz 1954. The long-term prognosis following myocardial infarction and some factors which affect it. Circulation 9:321-334. Danowski, TS 1957. Diabetes Mellitus with Emphasis on Children and Young Adults. Baltimore, The Williams & Wilkins Company, pp 1-510. Danowski, TS, NR Limaye, RE Cohn, BJ Grimes, JV Narduzzi et al. 1966. Sex distribution and frequency of diabetic concomitants or complications. Diabetes 15:507-510. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. The Overall Problem 29 Davis, MD 1965. Vitreous contraction in proliferative diabetic retinopathy. Arch Ophthal 84:741-751. Diabetes Source Book 1968. U. S. Department of Health, Education and Welfare, Public Health Service Publication No. 1168, revised 1968. Edmondson, HA, HE Martin, and N Evans 1947. Necrosis of renal papillae and acute pyelonephritis in diabetes mellitus. Arch Int Med 79:148-175. Ellenbert, M 1973. Current stuatus of diabetic neuropathy. Metabolism 22:657-662. Entmacher, PS, HF Root, and HH Marks 1964. Longevity of diabetic patients in recent years. Diabetes 13:373-377. Epstein, FH, LD Ostrander, BC Johnson, MW Payne, NS Hayner, et al. 1965. Epidemiological studies of cardiovascular disease in a total community--Tecumseh, Michigan. Ann Int Med 62:1170-1187. Falsetti, HL, and JD Schnatz 1970. Heart disease and diabetes mellitus. In Diabetes Mellitus: Theory and Practice. Editors, M Ellenberg, and H Rifkin. New York, McGraw-Hill Book Co., pp 870-889. Feldman, M, and M Feldman, Jr. 1954. The association of coronary occlusion and infarction with diabetes mellitus. A necropsy study. Am J Med Sci 228:53-56. Garcia, MJ, PM McNamara, T Gordon, and WB Kannell 1974. Morbidity and mortality in dia- betics in the Framingham population. Diabetes 23:105-111. Goldner, MG 1960. The fate of the second leg in the diabetic amputee. Diabetes 9:100-103. Goodman, JI, S Baumoel, L Siegfried, LJ Marcus, and S Wasserman 1953. The Diabetic Neuropathies, Springfield, Il1l., Charles C. Thomas. Grunnet, ML 1963. Cerebrovascular disease: Diabetes and cerebral atherosclerosis. Neurology 13:486-491. Halverstadt, DB, GW Leadbetter, and RA Field 1966. Pyelonephritis in the diabetic. JAMA 195:827-829. Hatfield, E. 1966. Extensified statistics on blindness and vision problems. In National Society for the Prevention of Blindness Fact Book. New York, The National Society for the Prevention of Blindness, Inc., pp 44-50. : Heptinstall, RH 1966. Pathology of the Kidney, lst edition, Boston, Little Brown and Co. : Hirohata, T, B MacMahon, and HF Root 1967. The natural history of diabetes I. Mortality. Diabetes 16:875-881. Imerslund, O 1959. The prognosis in diabetes with onset before age two. Acta Paediatrica 49:243-248. Kannel, WB, TR Dawber, A Kagan, N Revotskie, and J Stokes 1961. Factors of risk in the development of coronary heart disease--Six-year follow-up experience. Ann Int Med 55:33-50. Kaplan, MD, and AR Feinstein 1973. A critique of methods in reported studies of long-term vascular complications in patients with diabetes mellitus. Diabetes 22:160-174. Katz, LN, GY Mills, and F Cisneros 1949. Survival after recent myocardial infarction. Arch Int Med 84:305-320. Kessler, II 1971. Mortality experience of diabetic patients. A twenty-six year follow-up study. Am J Med 51:715-724. 30 48. 49. 50. 51. 52. 53. 54. 55. 56. 57 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. Diabetes Mellitus Kimmelstiel, P and WB Porter 1948. .Intercapillary glomerulosclerosis. New Eng J Med 238:876-879. Klimt, CR, GL Knatterud, CL Meinert, and TE Prout 1970. A study of the effects of hypoglycemia agents on vascular complications in patients with adult-onset diabetes I. Design, methods, and baseline results. Diabetes 19:747-783. Knowles, HC, GM Guest, J Lampe, M Kessler, and TG Skillman 1965. The course of juvenile diabetes treated with unmeasured diet. Diabetes 14:239-273. Knowles, HC Jr. 1971. Long-term juvenile diabetes treated with unmeasured diet. Trans Assoc Am Phys 84:95-101. Knowles, HC Jr. 1972. Glomerulosclerosis and Other Renal Disease in Juvenile Diabetes. Trans of the Second Beilinson Symposium on Juvenile Diabetes, Tel Aviv, Israel (In Press). Kramer, DW, and PK Perilstein 1958. Peripheral vascular complications in diabetes mellitus. A survey of 3,600 cases. Diabetes 7:384-387. Leading Article 1971. Coxsackie viruses and diabetes (editorial). Lancet 1:804. Leopold, IH, and TW Lieberman 1970. The eye and diabetes mellitus. In Diabetes Mellitus: Theory and Practice. Editors, M Ellenberg, and H Rifkin. New York, McGraw-Hill Book Co., Pp 796-821. Liebow, IM, HK Hellerstein, and M Miller 1955. Arteriosclerotic heart diesase in diabetes mellitus: A clinical study of 383 patients. Am J Med 18:438-447. Liebow, IM, VA Newill, and R Oseasolin 1964. Incidence of ischemic heart disease in a group of diabetic women. Am J Med Sci 248:403-407. Mayne, N 1965. Neuropathy in the diabetic and nondiabetic populations. Lancet 2:1313-1316. McGuinness, R 1967. Association of diabetes and cataract. Brit Med J 2:416-418. Miller, M, and ME Black 1940. Concurrent problems: Pregnancy. In Diabetes Mellitus: Diagnosis and Treatment, Vol. I. Editor, TS Danowski. New York, American Diabetes Association, pp 137-144. Moorhead, HB, et al. 1969. Causes of Blindness in 4,965 Persons Whom 14 States Added to Their MRA Registers in 1966. Proceedings of the 1968 Conference of the Model Reporting Area for Blindness, Washington, D. C., USPHS. National Center for Health Statistics 1964. Glucose Tolerance of Adults, United States 1960-1962. Public Health Service, Publ. No. 1000, Series 11, No. 2, Washington, D. C. National Center for Health Statistics 1967. Characteristics of Persons with Diabetes, United States, July 1964-June 1965. Public Health Service Publication No. 1000, Series 10, No. 40. National Center for Health Statistics 1967. Unpublished material, Health Interview Survey. National Center for Health Statistics 1973. Monthly Vital Statistics Report, Hospital Dis- charge Survey, Vol. 22, No. 6, Supplement. National Center for Health Statistics 1974. Monthly Vital Statistics Report, Provisional Statistics, Annual Summary for the United States, 1973. Vol. 22, No. 13. National Center for Health Statistics 1974. Mortality Trends for Leading Causes of Death, United States, 1950-1969. Public Health Service Publication No. 1000, Series 20, No. 16. Oakley, W 1965. The Treatment of Pregnancy in Diabetes Mellitus. In On the Nature and Treatment of Diabetes. Editors, BS Liebel, and GA Wrenshall. Publ. No. 84, Excerpts Medical Foundation, pp 673-678. 69. 70. 71. 72. 73. 74. 75. 76. 27. 78. 789. 80. 81. 32. 83. 84. 85. 86. 87. 88. 89. 90 The Overall Problem 31 Ostrander, LD Jr., T Francis, NS Hayner, MO Kjelsberg, FH Epstein 1965. The relationship of cardiovascular disease to hyperglycemia. Ann Int Med 62:1188-1198. O'Sullivan, JB 1961. Gestational diabetes. Unsuspected, asymptomatic diabetes in pregnancy. New Eng J Med 264:1082-1085. O'Sullivan, JB, and RF Williams 1966. Early diabetes mellitus in perspective--A population study in Sudbury, Massachusetts, JAMA 198:111-114. Partamian, JO, and RF Bradley 1965. Acute myocardial infarction in 258 cases of diabetes. Immediate mortality and five-year survival. New Eng J Med 273:455-461. Pedersen, J 1967. The Pregnant Diabetic and Her Newborn. Scandinavian Univ. Books, Munsgaard, Copenhagen, Denmark. Pell, S, and A D'Alonzo 1967. Some aspects of hypertension in diabetes mellitus. JAMA 202:104-110. Pell, S, and A D'Alonzo 1970. Factors associated with long-term survival of diabetics. JAMA 214:1833-1840. Rifkin, H, and J Berkman 1970. Diabetes and the Kidney. In Diabetes Mellitus: Theory and Practice. Editors, M Ellenbert, and H Rifkin. New York, McGraw-Hill Book Co., pp 848-869. Robbins, SL, and AW Tucker 1944. The cause of death in diabetes. A Report of 307 Autopsied Cases. New Eng J Med 231:865-868. Robinson, JW 1952. Coronary Thrombosis in Diabetes Mellitus: Analysis of 54 Cases. New Eng J Med 246:332-335. Root, HF, S Mirsky, and J Ditzel 1959. Proliferative Retinopathy in Diabetes Mellitus. Re- view of Eight Hundred Forty-Seven Cases. JAMA 169:903-909. Santen, RJ, PW Willis, and SS Fajans 1972. Atheroslerosis in Diabetes Mellitus: Arch Int Med 130:833-843. Stearns, S, MJ Schlesinger, and A Rudy 1947. Incidence and Clinical Significance of Coro- nary Artery Disease in Diabetes Mellitus. Arch Int Med 80:463-474. Tattersall, RB, and DA Pyke 1972. Diabetes in Identical Twins. Lancet 2:1120-1125. Thomsen, AC 1965. The Kidney and Diabetes Mellitus. Copenhagen, Munksgaard. Waddell, WR, and RA Field 1960. Carbohydrate Metabolism in Atherosclerosis. Metabolism 9:800-806. Waite, JH, and WP Beetham 1935. The visual mechanism in diabetes mellitus (A comparative study of 2,002 diabetics, and 457 nondiabetics for control). New Eng J Med 212:429-443. Warren, S, PM LeCompte, and MA Legg 1966. The Pathology of Diabetes Mellitus, 4th edition, Philadelphia, Lea &§ Febiger. West, KM, and JM Kalbfleisch 1971. Influence of nutritional factors on prevalence of dia- betes. Diabetes 20:99-108. White, P 1956. Natural course and prognosis of juvenile diabetes. Diabetes 5:445-450. White, P 1971. Pregnancy and Diabetes. In Joslin's Diabetes Mellitus, 11th edition. Editors A Marble, P White, RF Bradley, and LP Krall. Philadelphia, Lea § Febiger, pp 581- 598. . Whitehouse, FW, C Jurgensen, and MA Block 1968. The later life of the diabetic amputee. or Another look at fate of the second leg. Diabetes 17:520-521. 32 91. 92. 93% Diabetes Mellitus Wilkerson, HLC, and LP Krall 1947. Diabetes in a New England Town. A study of 3,516 persons in Oxford, Massachusetts. JAMA 135:209-216. Wilkerson, HLC, and JB O'Sullivan 1963. Study of glucose tolerance and screening criteria in 752 unselected pregnancies. Diabetes 12:313-318. World Almanac 1968. New York, Newspaper Enterprises Association, Inc. ECONOMIC IMPACT OF DIABETES Paul S. Entmach It is extremely difficult to estimate the full economic impact of diabetes, and the data that follow must be considered as significantly underestimating the situation. In developing the cost estimates relating to both morbidity and mortality, only statistics reported for diabetes as a primary cause were considered. Due to the nature of the disease, however, many of the complications, such as blindness and vascular disease, may lead to disability (and in the case of vascular disease to death), and the complication will be listed as causative with no reference to diabetes. Also, the impact that diabetes has as a contributing factor leading to disability or death from unrelated diseases cannot be measured. In addition, it must be remembered that the economic costs are derived from imprecise estimates. Despite these limitations, the cost esti- mates are of value in depicting the enormous impact that diabetes has on our economy. The methodology used to develop the cost estimates has been described by the National Heart and Lung Institute Task Force on Arteriosclerosis and their Task Force on Respiratory Diseases. In aggregate, the total economic cost in 1973 due to diabetes was approximately four billion dollars. This figure is arrived at by combining the following: 1. Direct costs due to illness: Expenditures for prevention, detection, treatment, re- habilitation, research, training and capital investment in medical facilities. In terms of types of services or object of medical expenditure, direct costs include amounts spent for hospital and nursing home care, physicians' and other medical professional services, drugs, medical supplies, research, training and other nonpersonal services, according to D. P. Rice in Health Economics Series No. 6, 1966. 2. Indirect costs due to morbidity: Lost earnings associated with man-years lost to productivity due to illness and disability. 3. Indirect costs due to mortality: Discounted present value of lifetime earnings of persons who died. As shown in Table 1, the total estimated cost resulting from illness and mortality due to diabetes in 1973 was 4.015 billion dollars. The total cost attributed to morbidity was 2.630 billion dollars with about 1.7 billion dollars due to direct costs and almost one billion dollars due to indirect costs representing 112,000 man-years lost from work. The costs due to mortality were 1.385 billion dollars, representing the present value of the remaining earnings of the 38,208 persons who died from diabetes in 1973. Statistics dating back to 1973 are used in this presentation because that is the latest year for which data are available. It is apparent that this is another cause for understatement of the costs because there has been significant inflation in the intervening two years. An adjust- ment can be made for the inflationary factor in the following manner. From May 1973 to May 1975, there was a 22.1 percent increase in the medical care component of the Consumer Price Index. This increases the direct morbidity costs from 1.65 billion dollars to 2.015 billion dollars. During the same period the average hourly earnings of production workers and nonsupervisory employees in 33 34 Diabetes Mellitus the private economy rose 15.5 percent. This increases the indirect morbidity costs from 980 million dollars to 1.132 billion dollars, and the costs attributed to mortality from 1.385 billion dollars to 1.6 billion dollars. Therefore, when inflation is taken into account, the total estimated costs of morbidity and mortality due to diabetes increases from four billion dollars in 1973 to 4.747 billion dollars in 1975. TABLE 1. Estimated Economic Costs of Morbidity and Mortality, Number of Deaths, and Man-Years Lost Due to Diabetes, United States, 1973. Costs (millions of dollars) Total $4,015 Morbidity 2,630 Direct 1,650 Indirect 980 Mortality 1,385 Number of Deaths 38,208 Man-Years Lost (thousands) 112 Source: Tables 2, 4, and 6. MORBIDITY Direct Costs Estimated direct costs of morbidity due to diabetes in 1973 by type of expenditure are shown in Table 2. The expenditures due to diabetes for 1) hospital care, 2) physicians' ser- vices, 3) drugs, 4) nursing home care, and 5) medical professional services other than by phy- sicians and dentists, are derived from the aggregate national health expenditure for each of the five categories by allocating a certain percentage of the total cost as being due to diabetes. The aggregate national health expenditures by type of expenditure are prepared by the Social Security Administration, and those for 1973 are shown in Table 3. TABLE 2. Estimated Direct Costs of Morbidity Due to Diabetes by Type of Expen- diture, United States, 1973.1 Type of Amount Expenditure (in millions) Total $1,650 Hospital Care 8002 Physicians' Services 4003 Drugs 225" Nursing Home Care 185 Other Medical Professional Services 40 lExcludes expenditures for dentists' services, eyeglasses and appliances, prepayment and administration, government and other health services, research and medical facilities construction. 2Based on days of care in short-stay hospitals. 3Cost of patient visits to physicians. “Cost of patient visits to physicians for which drugs were prescribed. Source: Estimated by the Statistical Bureau of the Metropolitan Life Insurance Company. Economic Impact 35 TABLE 3. Aggregate National Health Expenditure by Type of Expenditure, United States, 1973. Amount’ Type of Expenditure (in millions) Total gir $99,069 Health Services and Supplies 92,327 Hospital care 38,270 Physicians' services 18,200 Drugs and drug sundries 9,300 Nursing home care 7,050 Other professional services 1,900 Dentists' services 5,970 Eyeglasses and appliances 2,091 Prepayment and administration 3,998 Government public health 1,905 Other health services 3,643 Research and Medical Facilities Construction 6,742 Research 2,484 Construction 4,258 Source: Social Security Administration: Research and Statistics Note No. 1-75, Table 2, National Health Expenditures, Calendar Years 1929-73, February 19, 1975, Hospital Care: Care of diabetic patients for short-term hospitalizations cost 800 million dollars in 1973, representing over 5.2 million days of hospitalization. The total number of days of inpatient hospital care was obtained from the Hospital Discharge Survey which was published in Monthly Vital Statistics Report, and the number due to diabetes was esti- mated from data supplied by the Hospital Discharge Survey Branch of the National Center for Health Statistics. . The Hospital Discharge Survey is a continuing probability sample of all short-term hospitals in the nation excluding military and Veterans Administration hospitals and hospital units of institutions. By definition, short-term means under 30 days average stay per discharge. Patients were tabulated according to the diagnosis listed first on the summary sheet of the patients' records. The percentage of total days of care provided to diabetic patients was applied to the amount of hospital care expenditures for all illnesses (from Table 3) to yield the estimated amount of hospital care expenditures due to diabetes. This method assumes that the estimated expenditures for hospital care are distributed by diagnosis similar to the distribution of hospital days of care by first-listed diagnosis. Physicians' Services: The estimated cost of physicians' treatment of diabetic patients in 1973 was 400 million dollars. This represents about 34 million visits by physicians. The National Disease and Therapeutic Index estimated the total number of patient visits to physicians in private practice and the number due to diabetes. Its estimate is based on a continuing study of private medical practice in the United States in which data are obtained from a representative sample of physicians who report case history information on private patients seen over a period of time. The percentage of the estimated total patient visits to physicians due to diabetes was applied to the total amount of expenditures for physi- cians' services (from Table 3) to yield the estimated amount of expenditures due to dia- betes. The assumption is that the estimated expenditures for physicians' services are distributed by diagnosis similar to the distribution of patient visits to physicians by diagnosis. 36 Diabetes Mellitus Drugs: In 1973 the expenditure for drugs by diabetic patients was 225 million dollars. Estimates of patient visits to physicians in which medication was prescribed were reported by the National Disease and Therapeutic Index for the year 1970. The percentage of total visits in which medication was prescribed that was due to diabetes was then applied to the total amount of expenditures for drugs in 1973 to yield the estimated amount of expenditures due to diabetes. Nursing Home Care: Nursing home expenditures in 1973 for the care of diabetic patients was 185 million dollars. A survey by the National Center for Health Statistics showed the prevalence and distribution of chronic conditions among residents of nursing and personal care homes. The estimated percentage of diabetes among these residents was applied to the total expenditures for nursing home care in 1973, and a crude total expenditure for nursing home care for diabetics was computed. Other Medical Professional Services: Expenditures for other medical professional services rendered to diabetics in 1973 were 40 million dollars. These expenses include the cost of services provided by other than physicians and dentists. The portion of costs in this category that was due to diabetes had to be estimated by applying the percentage of the com- bined expenditures for hospital care, physicians' services, drugs and nursing home care to the total expenditures for other medical professional services. Indirect Costs Productivity loss because of illness and disability from diabetes in 1973 cost 980 million dollars. This is shown in Table 4. The productivity loss is measured in terms of man-years of work lost, and the economic value of the associated lost earnings. In 1973 diabetes caused 111,600 man-years lost from work. The estimated losses were computed for three population groups. TABLE 4. Estimated Man-Years Lost from Work and Keeping House and Indirect Costs of Morbidity Due to Diabetes, United States, 1973. Man-Years Lost Amount (in thousands) (in millions) Total 111.6 980 Currently employed persons 16.1 160 Women not in labor force who keep house 40.0 170 Persons unable to work 55.5 650 Source: Computed by the Statistical Bureau of the Metropolitan Life Insurance Company. Currently Employed Population: Among currently employed persons with diabetes there were 16,100 man-years lost from work with an associated loss of earnings of 160 million dollars. For the calendar year 1973 the estimated number of days lost from work associated with dia- betes was furnished by the Division of Health Interview Statistics of the National Center for Health Statistics. This is a continuing survey interviewing a representative sample of the nation's civilian noninstitutionalized population to determine the presence of acute and chronic conditions and the number of days of disability from these conditions. The estimated work loss days were divided by 245, the approximate number of working days in a year, to ob- tain the number of man-years lost. Mean annual earnings were applied to the man-years lost to obtain the indirect costs for the currently employed population prevented from working due Economic Impact 37 to diabetes. The mean earnings for 1973 were estimated by increasing the estimates by the National Heart and Lung Institute Task Force on Arteriosclerosis for 1964 by the ratio of the median income of men and women in 1973 to the corresponding median income in 1964 reported by the United States Department of Commerce, Bureau of Census. Average earnings for persons 14-44 years of age were assumed to be approximately the same as for persons aged 17-44. Housewives' Services: The estimated number of man-years lost from housework among diabetic women who keep house but who are not in the labor force was 40,000. This resulted in lost earnings of 170 million dollars. Housewives' services are estimated at the average earnings in 1973 for a domestic worker. This imputed value is clearly on the low side for it makes no allowance for the housewife's longer work week and takes no account of the size of the household cared for. Although the) economic contributions of housewives are not included in the national income accounts, according to the President's Committee on Heart Disease, Cancer, and Stroke, omitting the value of their services in the calculation of indirect costs dis- torts comparisons of costs among illnesses with different distributions by sex. Data from the National Health Interview Survey provided the estimated number of days of disability in bed due to diabetes for women, a figure which combines homemaker disability and the dis- ability of women in the labor force. The number of work-loss days was estimated by applying the proportion of all labor force members who are women to the total number of work-loss days for men and women combined obtained from the National Health Interview Survey. The number of female work-loss days was deducted from the bed disability days, and the resulting figure (adjusted bed disability days) was divided by 365 to obtain the estimated number of bed dis- ability years attributed to diabetes. Multiplying this figure by the percentage of women not in the labor force who keep house yields an estimate of the man-years lost for women who keep house. The imputed value of these man-years lost is obtained by multiplying the man- years lost by the average wage of a domestic worker in 1973 which was $4,190 per year, according to the United States Department of Commerce in a Survey of Current Business, 1974. Persons Unable To Work: Among persons unable to work there were 55,500 man-years of work. loss attributed to diabetes. This resulted in a production loss of 650 million dollars. The basic data were obtained from the Social Security Office of Research and Statistics. The proportion of total benefit awards during 1972 that were granted to applicants with diabetes was applied to the total number of disability beneficiaries at the end of 1973 to provide a conservative estimate of the man-years lost by those unable to work at all due to diabetes. Mean annual earnings were then applied to man-years lost to obtain the indirect costs. MORTALITY The indirect economic cost of mortality for diabetes to the nation is measured in terms of the discounted present value of lifetime earnings of persons who died. In 1973 this amounted to 1.385 billion dollars. The estimate was obtained by multiplying the 38,208 deaths from diabetes which occurred in 1973 by the expected value of future earnings with age and sex taken into account. Future earnings were discounted at 6 percent to take into account interest that could be earned in the interim with an adjustment made for increased productivity. The estimated pres- ent value of lifetime earnings by age and sex are shown in Table 5. These earnings were computed from discounted earnings available for 1964 from the National Heart and Lung Institute Task Force 38 Diabetes Mellitus on Arteriosclerosis which were increased by the ratio of the median income of men and women in 1973 to the median income in 1964, as reported by the United States Department of Commerce, Bureau ® of Census. TABLE 5. Estimated Present Value of Lifetime Earnings by Age and Sex, United States, 1973. Age Male Female Under 1 $55,208 $33,152 1- 4 63,054 38,046 5-9 84,667 50,618 10-14 113,526 67,842 15-19 146,759 84,398 20-24 174,753 92,081 25-29 186,391 92,141 30-34 183,246 90,677 35-39 171,424 87,460 40-44 152,541 82,974 45-49 127,602 76,232 50-54 100,337 68,017 55-59 73,281 59,415 60-64 43,097 49,308 65-69 21,926 39,993 70-74 18,263 32,580 75-79 13,981 23,958 80-84 8,119 13,004 85 and over 1,332 2,109 Source: Estimated by the Statistical Bureau of the Metropolitan Life Insurance Company The number of deaths from diabetes in 1973 is shown in Table 6. Also shown in this table are the indirect costs by age and sex which were derived by multiplying the number of deaths by the discounted earnings shown in Table 5. The total indirect costs due to mortality were 1.385 billion dollars with approximately 643 million dollars due to mortality among males and approximately 741 million dollars. due to female mortality. Economic Impact 39 TABLE 6. Number of Deaths and Estimated Indirect Costs Due to Diabetes, United States, 1973 Number of Deaths Indirect Costs (in thousands) Age Total Male Female Total Male Female Total 38,208 15,669 22,539 $1,384,639 $643,143 $741,496 Under 1 16 11 5 773 607 166 1- 4 14 9 5 757 567 190 5-9 22 8 14 1,386 677 709 10-14 52 19 33 4,396 2,157 2,259 15-19 66 33 33 7,628 4,843 2,785 20-24 126 64 62 16,893 11,184 5,709 25-29 243 143 100. 35,868 26,654 9,214 30-34 320 171 149 44,846 31,335 13,511 35-39 427 247 180 58,085 42,342 15,743 40-44 616 330 286 74,070 50,339 23,731 45-49 1,046 533 313 107,119 68,012 39,107 50-54 1,752 892 860 147,996 89,501 58,495 55-59 2,715 1,316 1,399 179,560 96,438 83,122 60-64 3,847 1,749 2,098 178,825 75,377 103,448 65-69 5,139 2,256 2,383 164,765 49,465 115,300 70-74 6,146 2,422 3,724 165,561 44,233 121,328 75-79 6,347 2,342 4,005 128,696 32,744 95,952 80-84 5,303 1,843 3,460 59,957 14,963 44,994 85 and over 4,008 1,280 2,728 7,458 1,705 5,753 Source: Deaths from the National Center for Health Statistics, Division of Vital Statistics, and Table 5. Acknowledgments: Appreciation is expressed to Mr. Stanley Kranczer, Supervising Research Assistant, and Ms. Lena Bittenson, Statistical Analyst of the Statistical Bureau, Metropolitan Life Insurance Company, for their assistance in preparation of this material. REFERENCES 1. Lea Incorporated, National Disease and Therapeutic Index 1969. Leading diagnosis in physician visits. Review, June 1972. 2. National Center for Health Statistics, Division of Health Interview Statistics Undated. Unpublished information obtained by personal communication. 3. National Center for Health Statistics, Division of Health Resource Statistics, Hospital Discharge Survey Branch Undated. Unpublished information obtained by personal communication. 4, National Center for Health Statistics, Hospital Discharge Survey 1969. Utilization of short-stay hospitals--summary of non-medical statistics. United States. Monthly Vital Statistics Report, Vol 21, No 6, Supplement. 5. National Center for Health Statistics 1967. Prevalence of chronic conditions and impairments among residents of nursing and personal care homes. United States, May-June 1964, Washington, DC, Government Printing Office, PHS Publication No 1000, Ser 12, No 8. 6. National Heart and Lung Institute Task Force on Arteriosclerosis 1971. Arteriosclerosis, Vol II. Washington, DC, Government Printing Office, DHEW Publication No (NIH) 72-219. 7. National Heart and Lung Institute Task Force on Respiratory Diseases 1972. Respiratory Disease. Washington, DC, Government Printing Office, DHEW Publication No (NIH) 73-432. 8. The President's Commission on Heart Disease, Cancer, and Stroke 1965. A National Program to Conquer Heart Disease, Cancer, and Stroke, Vol 2. Washington, DC, Government Printing Office. 40 Diabetes Mellitus 10. 11. 12. 13. 14. Rice, DP 1966. Estimating the cost of illness. Health Economics Series No 6, USPHS Publication No 947-6. Social Security Administration 1972. Social Security Disability Applicant Statistics, 1968. Washington, DC, Government Printing Office. Social Security Bulletin 33 1970. Annual Statistical Supplement, 1969, p 75. United States Department of Commerce, Bureau of the Census Undated. Income in 1964 of families and persons in the United States. Current Population Reports, Ser P-60, No 47. United States Department of Commerce, Bureau of the Census Undated. Income in 1969 of families and persons in the United States. Current Population Reports, Ser P-60, No 75. United States Department of Commerce 1970. Survey of current business 50:40. IMPROVING THE ORGANIZATION OF CARE FOR THE CHRONICALLY ILL Leona V. Miller, Jack Goldstein, and John W. Runyan, Jr. BACKGROUND Nearly half the population of the United States suffers from one or more chronic disease (8). This segment of the population is estimated to account for 80 percent of all health care services delivered. If health care costs are to be reduced in any meaningful way, the major stress will have to be in improving health care delivery to the chronically ill. New methods of health care delivery will have to be developed. Pilot programs with representative chronically ill groups will need to be evaluated before widespread changes can be carried out with the total chronically ill population. Perhaps one of the most representative chronically ill groups are diabetics, as there is hardly an organ system which diabetes may not affect. Organizing systems to deliver care to dia- betic patients incorporate, in microcosm, all the problems found in developing systems to care for chronically ill patients in general. For example, the vast majority of admissions to the Diabetes Service at the Los Angeles County University of Southern California Medical Center are not primarily for problems of diabetes management, but rather for other chronic conditions often found among diabetics, such as heart disease, renal disease, pulmonary disease, and neurological problems among others. The probability that a diabetic will be hospitalized in any one year is 67 percent. The ic spends an average of 5.4 days in the hospital each year and loses an average of 15.4 per year. Contrast these statistics with that of the nondiabetic. The likelihood that | be hospitalized is but 12 percent annually. He spends an average of one day in the hos- pital each year and loses an average of only 5.7 work days (U.S. National Center for Health Sta= tistics). The inability of the health system to respond fully to the diabetic patient's needs affects his life in a variety of ways. Employment may present a challenge to him. Insurance companies, although more liberal in recent years, do exclude health benefits to diabetics (and a variety of other chronic disease groups) under a ''preexisting condition" clause which reduces benefits to diabetic subscribers for the first 2 years of a health insurance policy and in many cases provides inadequate benefits thereafter. The result is that diabetes is not merely a medical diagnosis, but it also inhibits the dia- betic in his ability to obtain a better job and often jeopardizes his financial position. The basic problem is to find a way of delivering care to the chronically ill population so that quality of continuing care is more accessible, which in turn may reduce both morbidity and the costs of providing their care. Once a solution is designed for this basic problem, the effect of the collateral problems will be minimized. CURRENT STATE OF KNOWLEDGE A number of investigators have acknowledged that the prime utilizers of health care services in this country and elsewhere, are the chronically ill (1). Little, however, has been accom- plished in designing health systems to deal specifically with this group. Because of the 41 42 Diabetes Mellitus economic and political orientation, health delivery systems have developed in a fragmented manner which often do not necessarily meet the needs of the chronically ill. However, there are two models which attempt to deliver care to the chronically ill and the diabetic patient in an organized manner. The first of these was initiated over 11 years ago in Memphis and Shelby counties (7,6). From the start in 1963, this program has been a joint effort of the local Health Department, the City of Memphis Hospital, and the University of Tennessee College of Medicine. This program's concept of the care of patients with chronic diseases includes two essential ingredients: (a) geographic decentralization of the health care facilities for ready accessi- bility to the patients, and (b) the use of specially trained nurses guided by protocols with physician and medical center back-up for patient management to these facilities. There are currently 18 decentralized clinics. Ten are in urban areas and eight in the semi- rural areas; these serve a population of 225,000 out of the county's population of 750,000. Four of these are multipurpose neighborhood health centers, while the remainder are small installa- tions, called "satellites," located in community centers, public schools, and housing develop- ments. A mobile bus extends medical care to more inaccessible places and remains for the entire day in each of five locations. ‘In 1973 over 40,000 visits were made to these facilities by the more than 9,000 chronic disease patients who are in the program. This includes approximately 3,500 diabetic patients. The equivalent of 20 full-time, specially trained nurses provided the maintenance care. The low missed-appointment rate of 5 to 8 percent is a reflection of the easy accessibility of the facilities, the devotion of the nurses administering the care, the excel- lence of the follow-up system, and the reduced costs to the patients. It contrasts to the 25-50 percent missed-appointment rate in outpatient clinics of the City Hospital. The flow of the patients from the decentralized locations to the secondary care facilities and back again is expedited by a liaison service. This service is composed of two nurses, one \representing the hospital and the other the Health Department. The consulting physicians in the University Medical Center serve as the medical back-up for the liaison nurses and the nurses in the neighborhood clinics and home. These consulting physicians have developed the medical pro- tocols and policies for the program with the cooperation of the nurses delivering the direct care. : Some observations have been made on a subset of the chronic disease population involving 1,006 patients transferred from the city hospital to the decentralized facilities for maintenance care (6). In this subset, 555 patients had diabetes mellitus. Ambulant care before transfer was primarily in the hospital clinics and in the hospital emergency ward with a combined total of 8,124 visits per 1000 patients per year to these facilities. After transfer there was a shift, as expected, to the decentralized clinics. Combined outpatient department and emergency ward visits fell to 2,585. Some 166 home visits were also made on this subset, an important index of the continuum of care provided and the increased patient contact with health professionals after transfer. The cost of ambulant care received in the decentralized locations was one-fifth the cost of comparable care rendered in medical center facilities. A visit to a neighborhood clinic cost $4 compared to $20 for a hospital outpatient visit and $35 for an emergency ward visit. When the pattern of hospitalization for the 2-year period before and after transfer was Improving the Organization of Care for the Chronically Ill 43. examined for those with diabetes, a 49 percent reduction in hospital days and a 42 percent reduction in hospital discharges were found (Table 1). The decrease in the hospital days for diabetic acidosis, serious infections, and problems related to peripheral vascular disease, the preventable complications of diabetes, accounted for the larger share of these reductions. The hospital days required for the renal and vascular complications of diabetes showed, on the other hand, a slight increase indicating the limitations of continuing care in altering the course of these progressive, degenerative complications. The increase of 20 percent in this group was due primarily to readmission to patients with advanced and progressive renal insufficiency frequently with multiple complications. TABLE 1. Memphis and Shelby County Continuing Care Program. Hospital Utilization! Diabetic Patients "Per Thousand Patients Per Year Hospital Days Hospital Average Discharges Medical Surgical Total Per Admission Before Transfer? 211 1297 2022 3319 15.7 After Transfer? 123 746 934 1680 13.6 Reduction 88 (42%) 551 1088 1639 (49%) Ipata based on a survey of 555 diabetic patients transferred to decentralized facilities for con- tinuing care between September 1, 1969 and August 31, 1970. 2Data based upon hospitalizations two years before transfer and two years after transfer. The 49 percent decrease in the total costs of the medical care provided in this program as compared to the costs before transfer is due to the use of the less expensive ambulant services and the decrease in hospitalizations (Table 2). TABLE 2. Memphis and Shelby County Continuing Care Program. Medical Care Costs Estimates Per 1000 patients per year Before Transfer After Transfer Ambulant and Home Care? Decentralized Clinic Visits @ $4 $ 24,732 Hospital Out-patient Visits @ $20 $ 157,380 50,820 Emergency Ward Visit @ $35 9,025 1,540 Home Visit @ $12 - 1,992 Laboratory @ $24.96 24,960! 24,960 Medications @ $54.75 54,7501 54,750 § 246,115 $158,794 Hospital Care3 $ 331,900 $168,000 Total Costs $ 578,015 $326,794 lExact figures unobtainable. 2Based upon data from 1006 chronic disease patients including 555 diabetic patients before and after transfer to decentralized facilities. 3Based upon data from 555 diabetics before and after transfer to decentralized facilities. 44 Diabetes Mellitus Past studies using health aides, medical students, and nurses, both in the clinic and in the home, indicate favorable acceptance by the overwhelming majority of the patients of this kind of personnel acting in a professional role. In this study, 98 percent of the 662 patients interviewed preferred this care to other choices. Sixty-four percent considered their general state of health had improved since transfer; only 5 percent felt that their health status had worsened. The effectiveness of the program for continuing care of patients with diabetes mellitus and chronic diseases is reflected in (a) the highly significant reduction in visits to the clinics and emergency room, (b) the reduction of admissions to the wards of the City Hospital for the preventable complications of diabetes, (c) the approximately 44 percent reduction in cost of continuing care as compared to the cost by conventional methods in use in the same community, and (d) the enthusiastic acceptance of the system by the patients themselves. While the access to health care has been improved and costs reduced in the Memphis program, financial resources are insufficient to make this system more responsive to patients' needs. A patient needing medical advice or care when the neighborhood health clinic is closed can satisfy his need only by appearing at the emergency room. This tends to disrupt the continuity of care so necessary in caring for the chronically ill. The system in use on the Diabetes Service at the Los Angeles County--University of Southern California Medical Center (LAC-USC) has attempted to solve this problem in another manner. Here too, Nurse Clinical Practitiongrs are employed to provide care to patients. However, this care is provided within the confines of the Medical Center. This tends to make the care somewhat less accessible than that afforded by the Memphis group. However it tries to improve continuity of care by: 1. furnishing a 7-day-a-week, 24-hour hotline telephone service for its patients; 2. offering medical service at all times either on a scheduled basis in the outpatient de- partment or on an unscheduled basis in a special walk-in clinic maintained by the Dia- betes Service. Patients coming to this clinic are evaluated and either cared for on an ambulatory basis, or if diagnostic tests warrent, admitted. Unnecessary admissions have been virtually eliminated by the patients being screened in this Diabetic Walk-In Clinic. 3. maintaining up-to-date records of all 8,000 patients followed by the Diabetes Service in specially formulated, condensed ledgers retained on the Diabetes Service. These ledgers are part of an overall information system that not only permits the rapid information retrieval necessary for providing patient care but also permits computerized review of each physician's performance in health care delivery, the performance of the entire Dia- betes Service, as well as epidemiological studies of the inpatient population. A more complete description of the details of operation of this system is contained in the literature (4). The effect of this system on reducing costs has been extremely impressive (Table 3). Since this system has been in operation, the average number of days spent by a clinic patient in the hospital annually has decreased from 5.6 to 1.7. This figure is not significantly different from the 1-day average of a randomly ill population. It has meant that the probability of a clinic patient being hospitalized in any one year has decreased from .67 to .21. Total costs of hospitalization during the system's first full year of operation of the Diabetes Service have been reduced by over $3 million. This reduction of hospitalization has resulted in annual savings of $588 per clinic patient. Improving the Organization of Care for the Chronically Ill 45 TABLE 3. Los Angeles County Hospital.? Diabetic Hospital Care Per Year Hospital Discharges Hospital Days Hospital Cost per 1000 pts. Average per pt. per pt. 1968 670 5.60 (5.4)! $ 840 19702 208 1.74 $ 261 Reduction 462 (70%) 3.86 (70%) $ 579 INational average. 2After introduction of communication and information system to improve patient education. Unfortunately, health systems for the chronically ill, such as that in Memphis and Shelby counties and the Los Angeles County--University of Southern California Medical Center, are not common. Incentives to increase the efficiency and reduce the cost of health care delivery are lacking. Some argue that a system of prepaid health care would change these incentives and would stress efficiency and cost control. However, it does not necessarily follow that the health status of the subscribers would also be properly maintained, so any such system would need ap- propriate surveillance. .No single change in the financing of health care can by itself improve health care delivery, control costs, or help in the redistribution of health professionals and facilities. It is the health care organization that is deficient, and this is where research is needed if the system is to become more appropriate and responsive to the needs of our popula- tion. Research is required on how best to focus resources on needs and on how to provide the ap- propriate financial and other incentives necessary. If incentives are appropriate, income to the health system would be based not on the output of the system, but on the effectiveness of the system. It would be judged on how well a system maintained the health of its population relative to standards established for that population based on disease distribution, age, sex, education, financial status, and other variables that may affect health status. For this to be accomplished, a major development effort has to be initiated in improving ac- cess to information about the health system. Research into understanding the complexity of the present health system and improving it is impeded by an obsolete information system relying prin- cipally on the medical record. Numerous articles have been printed on the deficiencies of the medical record (2,3,5,9), but insufficient stress has been placed on the inimical effect this type of documentation has on: delivering care to the patient; monitoring professional perfor- mance; maintaining organizational efficiency; and controlling costs. The use of medical records as the only source of information is effective primarily for the solo practitioner and small groups. It was never designed to be used by large organizations such as hospitals where the need for information contained in the medical record is needed by many diverse users for different purposes, often at the same time. A more automated, better organized system is needed that can deal with large data bases and deliver information more readily to the physician, his chief, the hospital administrator, the planner and the researcher. 46 Diabetes Mellitus To achieve this, research is needed in design and development of improved information systems to capture, store, retrieve, and manipulate data. A new type of health professional with strong technical background will have to be developed who can not only communicate with other health professionals but who can also use mature judgment in designing information systems much as an architect does in building homes. With the development of new information systems, research into introducing new systems of health care delivery and evaluating them objectively can be expedited. REFERENCES 1. Brook, RH, FA Appel, C Avery, M Orman, and RL Stevenson 1971. Effectiveness of inpatient follow-up care. N Engl J Med 285:1509-1514. 2. Day, E 1970. Automated health services--Reprogramming the doctors. Methods of information in medicine 9:116-121. 3. Last, JM 1970. Data collection requirements for patient management. Medical Care 8:159- 168, supplement. 4. Miller, LV, and J Goldstein 1972. More efficient care of diabetic patients in a county hospital setting. N Engl J Med 286:1388-1391. 5. Reid, MH 1972. Fishing through patient records. Medical Care 10:488-496. 6. Runyan, JW 1973. Decentralized medical care of chronic disease. Clinical Research 21:276. Paper to be published in Transactions of the Association of American Physicians, Vol. 83. 7. Runyan, JW, WE Phillips, O Herring, and L Campbell 1970. A program for the care of patients with chronic diseases. JAMA 211:476-679. 8. Somers, AMR 1971. Health care in transition: Directions for the future. Chicago Hospital Research and Educational Trust, p 20. 9. Tufo, HM, and JJ Speidel 1971. Problems with medical records. Medical Care 9:509-517. 10. United States National Center for Health Statistics, characteristics of persons with dia- betes: United States. July 1964 - June 1965. Publication No 1000, Series 10, No 40. Washington, DC, Government Printing Office, 1967, p 9. THE COMPUTER IN THE MANAGEMENT OF DIABETES Robert E. Bolinger The rapid progress in the application of the computer to hospital financial accounting sys- tems contrasts sharply with the relatively slow progress made in application of the computer to health care management. Many of the earlier attempts to apply the computer to health care manage- ment failed because of a lack of understanding between the data processing professionals and the health care professionals. A present more cautious attitude exists (3) which takes cognizance of three central problems. (a) A thorough analysis and understanding of the health care system must be feasible and occur before any attempt is made to apply the computer. (b) Facilities for on-line data storage must be extensive. Technology in this respect is advancing and promises relatively cheap on-line storage facilities. (c) Logical systems or software must be developed which readily handle clinical information. A number of systems of this type are currently being developed, such as HIPS and SETRAN (34), MUMPS (14), DATAFLOW (8), and others. The characteristics of the computer which render it uniquely useful in health care manage- ment include its capacity to store large amounts of data; to perform mathematical computations on the data which reveal trends, means, and variability; to carry out reproducible logical sequences based on either discrete or continuous data; and to handle multifactorial problems which are com- mon in the clinical setting. A major stricture often encountered in any computer application to health care has to do with the method of entering data, since a given clinical situation often presents with a large number of variables. Thus, the time required for a human to enter the con- dition of each clinical variable into a machine may nullify some of the advantages listed above. Diabetes presents problems of management both as an acute disorder and as a chronic dis- order, and it is in the latter situation where the potential of the computer to store large amounts of data and derive trends is particularly useful. Computer applications in the manage- ment of diabetes may be considered in three categories: record processing, simulation, and decision making at both the diagnostic and therapeutic levels. RECORD PROCESSING A familiar sight, particularly in a diabetic clinic, is a very thick chart written in a multitude of formats in which significant information is often buried in a myriad of apparently irrelevant data. A major thrust in health care research is directed at the development of technics * to reduce these data to a form which may be stored in a computer system, retrieved therefrom in a usable format, and statistically processed. Four types of clinical data that are essential are the updated historical and physical examination information, laboratory data, and an outline of problems and treatment. Such information may be stored as purely text information in a form not unlike that found on a regular hospital chart or as coded information. The latter has ob- vious advantages since it requires far less storage than text information does and can be readily retrieved on the basis of code. It is possible by the use of some extensions of any of the standard medical terminology coding systems (ICDA, H-ICDA, CMT, SNOPS) to handle most of the in- formation of the history, physical examination, and the problem list as coded information. Once 47 48 Diabetes Mellitus information becomes encoded, it becomes not only readily retrievable but also amenable to various forms of data processing for means, variabilities, and trends, so at the same time a data base for statistical and research purposes can be created. It is not surprising, in view of the many problems which are encountered in the long-term management of the diabetic patient, that there exists such a paucity of reliable statistical confirmation of supposedly accepted modes of prac- tice. Certain aspects of the medical history, particularly system review, past history, and family history are readily adaptable to computerized systems. Slack et al. (29) have initiated the work on computer-based histories, and a number of others have followed (7,16,22,25) The mode of entry of historical data may vary from a fixed set of questions which are asked the patient to a com- plicated set of branchings which, through direction from the computer carry the interrogative thread into the areas where emphasis is needed. Extensively branched histories have been de- veloped by Weed et al. (34), by Gottlieb (13), and Grossman et al. (16). From the standpoint of the diabetic patient, an extensively branched history would yield the largest amount of informa- tion, but as a practical compromise, certainly, a detailed system review screen with the addition of specific questions highlighting the development of the diabetic process would suffice. An effective input technique for this procedure is some type of optical scanning method whereby the marks indicated by the patient on a questionnaire can be directly read into a machine system. This minimizes the use of the computer time and allows the patient to pace himself according to his own ability. The entry of physical examination data requires some form of physician input and may require provision for branching. This would be true in the case of the diabetic where particu- lar emphasis would be needed on details about the skin, the retina and other features of the eye, the peripheral vascular system, the nervous system particularly related to diabetic neuropathy, etc. Although it might be argued that these features might be part of any physical examination, experience teaches us that they are frequently omitted. Therefore, the inclusion of these in some type of a structured physical examination, slanted particularly toward the diabetic patient, should be of particular advantage in obtaining a reliable data base with respect to these features. The technology of the processing of laboratory data is much better developed and has indeed become a standard part of many health care systems throughout the country. The features of laboratory data processing which are of particular interest in the management of the diabetic pa- tient, in addition to the simple storage and retrieval of the raw data itself, are the reduction of data and the derivation and computation of trends which can serve as essential guidelines (19) in the management of the patient either by the physician or by other computer programs. The storage and retrieval of problem lists (21,35)can be of particular importance in the management of long-° term chronic illness such as diabetes, since it highlights the areas which need attention. The use of the computer in the storage of problem lists is particularly convenient because the lists can be much more easily updated and kept current. Furthermore, the problem list data can be coded according to standard coding procedures and therefore become readily retrievable for statisti- cal purposes in the study of a diabetic population. The storage and processing of treatment and other management procedures carried out in the patient are particularly important. The patient may be receiving treatment for several different problems besides the diabetes, and often the total management picture of the patient tends to be forgotten by a preoccupation with insulin dosage and medications and other procedures which may The Computer in the Management of Diabetes 49 actually be influencing the diabetic status. A readily retrievable display of all current treat- ment should be a particular advantage in the management of this type of patient as, also, should be a record of results of past medications. Trends in the slow panorama of events representing the development of diabetic complications, such as retinopathy (27), can be successfully followed by the computer. Disease registry systems are particularly important in the management of chronic illnesses and many diabetic clinics have registry information over a long period of time. Such registry systems particularly permit the storage of prospective data, and in such notable instances as the UGDP (1970), the interdigitation of the computer facilities with a registry system provided a tool for the integration of a mass of information. In particular, registry information be- comes a data bank which can be processed and reduced to computable forms and used in the on-line management of the diabetic patient by providing the current updated statistics and probabilities necessary to make decisions as described below. At the same time, if the registry system is interlocked with the on-line management system, it can be continually and regularly updated by the introduction of the coded information at each point in the patient's course. The effectiveness of the computer-based record in the management of diabetes has been demon- strated by Thomas and Dobson (32) at the Hermann Hospital in Houston, Texas. A system of initial data base entry by punched cards is followed by updating the computer record at each subsequent visit. Summary printouts of the data in problem-oriented format are made available at each visit and included in the permanent record. SIMULATION A second area in which the computer has contributed to diabetic management is in the process of simulation. The complex system involved in the regulation of blood glucose can be simulated in the computer (6,23), thus providing a model for research on the development of automatic in- sulin release devices. Srinivasan et al. (31) developed, by the use of the computer, a model of differential equations to fit the experimental data derived from glucose changes. Grodsky et al. (15) published data using the perfused pancreas and studied insulin secretion, thereby isolating data from a single component of the system. Gatewood et al. (9) and Gatewood et al. (10) proposed a simplified model for the regulation of glucose in diabetics. They were able to show with the computer that most of the variations in the blood sugar could be explained by the insulin level alone. DECISION MAKING Clinical decision making has been studied in both the diagnostic and the therapeutic setting. From the diagnostic standpoint, the computer may be valuable first in examining the data from a large patient population and establishing the criteria for the detection and diagnosis of diabetes. This has been extensively used in multiphasic health screening. Hecker (17) noted that 10 percent of the population have 50 percent of the disease and emphasized the importance of setting up specific markers for the detection of these individuals in a screen. Gleser and Collen (11) have developed a model for selecting the significant variable to be tested for in multiphasic health screening and diabetes detection. This is particularly important in diabetes detection where historical factors, nutritional factors, and certain laboratory data may be a clue before specific symptoms of the disease appear. 50 Diabetes Mellitus Karp et al. (24) and Silcock et al. (28) used a computer to analyze the glucose insulin curves during the glucose tolerance test to identify the parameters that could differentiate accurately between normal and abnormal groups. In connection with the multiphasic health screen- ing, Himbert et al. (18) and Bastenie et al. (2) used the computer to study the importance of diabetes among the risk factors encountered in the screening of cardiovascular disease. The use of the computer as a component of the medical management process is even less developed. Isolated efforts in various fields have applied the computer in part to management de- cisions in respiratory disease (26) and acid base balance (12). The process of medical decision broadly encompasses--first, the collection of data; second, the processing of the data to arrive at a diagnosis; and finally, the prescription of a plan of action designed to correct the problem. In practice these steps are often telescoped and tend more to approximate a guidance system in which data collection serves as a continuous feedback for modification of diagnosis and for the evolution of a plan of action based on the outcome of the previous action. The central control- ling or stopping function for this process is the solution of the patient's problems. Often in clinical practice there may be among the alternatives decisions which are potentially hazardous or even fatal to the patient. Therefore, in each step of a decision process, the logic must be so structured that hazardous outcomes are given a probablility of zero. A large part of the educa- tion of a physician is spent in developing skill in this process and, in the case of diabetes, his activities often assume a repetitive pattern, some aspects of which can be easily assumed by a computer. In this clinic Bolinger et al. (4), Bolinger et al. (5) have tested the feasibility of computerizing some of the aspects of the management of diabetes. In this study it was as- sumed that the diagnosis of diabetes had been established. In the diabetic clinic of the Kansas University Medical Center, the variables which are usually queried in the day-to-day management of the diabetic patient were isolated and a system devised whereby these variables could be entered into the system through a teleterminal and submitted to a series of Boolian operators, leading to the prescription of insulin and diet, as well as to other warning flags in patient management. Provision was made for prescription of insulin dosage in short acting, intermediate acting, and long acting, and for insulin administration at morning, noon, evening, and midnight. Sox et al. (30) have emphasized the possibility of broadening the use of physician assistants in the management of disease by the use of computerized algorithms. It had been possible in his studies to essentially relieve the physician's assistant of any complicated decision process and utilize him strictly as the data-gathering device and as an instrument in prescribing, after the logical work was accomplished by the computer. In Phase II of the diabetes management project at the Kansas University Medical Center, a nurse clinician was used in this capacity, supported by the logical processing of the computer as an on-line decision-making resource, utilizing informa- tion which she gathered at that time. The output of the computer was compared with the recom- mendations of a diabetologist, and the correlation between the computer recommendations and those of the diabetologist for insulin dosage are shown in Table 1. The three phases of the study indicated in the table are based upon versions of the program which in Phase I was overly simpli- fied and therefore not very flexible. Phase II was characterized by the intervention of the nurse clinician, and Phase III represented a final revision of the program, including inpatients. The correlations between the computer recommendations and those of the physician are generally satisfactory, and it is thought at the present that hazardous-type decisions which might lead to The Computer in the Management of Diabetes 51 serious hypoglycemia or a diabetic ketoacidosis have been eliminated. The complexity of the de- cision process as monitored by a trace program was reflected in the number of times the program had to branch and was for insulin-taking diabetics and noninsulin requiring diabetics, 14 and 7 times respectively, indicating the greater simplicity of the logical process in the noninsulin- taking diabetic. With more data of this type, it should be possible to decide upon the level of complexity of a decision which paramedical personnel might be able to handle without the support of a computerized algorithm. This type of analysis of a computerized system which tends to quanti- tate the complexity of a decision has broad implications in the allocation of responsibility in the clinical process. TABLE 1. Correlation of Insulin Recommended by Computer vs. Physician Insulin Tr P Phase I (N = 14) AM Lente .949 <.001 PM Lente .000 >.05 AM Regular .686 <.01 PM Regular .000 >+05 Total +983 <.001 Phase II (N = 49) AM Lente .918 <.001 PM Lente .790 <.001 AM Regular .955 <.001 PM Regular .649 <.001 Total .864 <.001 Phase III (N = 99) AM Lente .906 <.001 PM Lente .831 <.001 AM Regular .878 <.001 PM Regular «2 77 <.001 Total .928 <.001 Coefficient of correction Statistical probability using Students t test Number of cases ZoH noun Reprinted with permission from Diabetes 22:480-484 (June 1973). Journal of the American Diabetes Association. The use of a logical model as described by Bartholomay (1) does not appear necessary at the level of resolution ordinarily encountered in diabetic management. Instead, a system of sorting, as shown in Fig. 1, was used with a system of filters which assigned patients either to one group or another at each logical decision point. The model also provided for transfer of patients from the insulin-taking to the noninsulin-taking group and vice versa, depending upon changing conditions. It is concluded that this type of computer program did demonstrate the feasibility of rela- tively complicated algorithms in the management of some phases of diabetes. It is also concluded that such a system could stand alone without fairly extensive systems of a similar type handling many of the common intercurrent diseases which occur in connection with diabetes, such as hyper- tension, heart failure, infections, and some of the other endocrinopathies. 52 Diabetes Mellitus DECISION FLOW CHART DIABETIC PATIENT TAKING NOT TAKING INSULIN == INSULIN 7 il - g WELL MAL- | ~ MAL- WELL REGULATED REGULATED|~ — — __ __ REGULATED REGULATED ze | Ne ie | re — ADJUST CHANGE RE- ON ORAL DIET DIET INSULIN DISTRIBUTE HYPOGLYCEMIC |~——| ALONE DOSAGE INSULIN AGENTS | CHANGE CHANGE ADJUST TYPE DOSE DIET FIGURE 1. The process of grouping which is carried out through the algorithm of the program is shown. At the top, a heterogeneous group of diabetics is postulated and, at the bottom, the sorting of these patients into treatment groups results from the sorting. RELATION TO THE HEALTH CARE SYSTEMS An idealized summary of the possibilities for computer assistance in the management of diabetes is shown in Fig. 2. The proposed allocation of functions consigns activities to five areas; i.e., the physician, the paramedic, the man-machine interface, the computer, and its COMPUTER BASED DIABETIC MANAGEMENT PHYSICIAN PARA-MEDIC COMPUTER COMPUTER STORAGE INTERFACE PATIENT ADMISSION I Junsirnaze HEALTH ] ouipe muLTIPHASIC SCREENING HEALTH SCREENING AUTOMATED OPTICAL SCANNER DATA HISTORY BASE CRT STRUCTURED CARDS CODE AND STORE PHYSICAL TELE TYPE DATA BASE EXAMINATION 1 HISTORY OF LAB. DATA PRESENT ILLNESS $ 0 DIAGNOSIS | ririre==] PROBLE INITIAL ORDERS le EDUCATION FOLLOW-UP VISITS NEEDS TO SE PHYSICIAN CURRENT STATUS SUMMARY OUTLINED DATA BASE UPDATED FLAG WARNINGS PHYSICIAN VISIT ASSEMBLE CURRENT STATUS SUMMARY R DECISIONS Te , FIGURE 2. A proposed allocation of the informational work is shown for various phases of the management of the diabetic patient. Areas of allocation include physician, paramedical personnel, computer-man interface, computer and infor- mation storage. PROCESS REPORTS The Computer in the Management of Diabetes 53 slow and fast storage capabilities. The frequent exchange of information between these different components emphasizes the necessity of an integrated plan of development and design. The feasi- bility of the application of each of the components of this scheme have been demonstrated, but in no place is it totally operative. The greatest progress in terms of a functioning system is de- scribed by Thomas and Dobson (32). An important area for further research is indicated to deter- mine the functioning specifications of each of these components needed to obtain optimal integrated performance. Any such testing should be undertaken only with a clear definition of the response variables, such that a clear distinction between present and proposed systems can be demonstrated. The response variables to be evaluated should include utilization of physician time vs. para- medic time, convenience to patients, impact on the patients' attitudes regarding the patient- physician relationship, effect on long-term cooperation by the patient, effect on patient educa- tion about his disease, long-term morbidity and mortality statistics particularly emphasizing the role of the computer in possible prevention of diabetic complications, costs and convenience of various options available at the man-machine interface, computer hardware and mass data storage systems. With respect to data storage, agreement is needed among authorities in the field of diabetes as to exactly what data should be available as on-line storage for diabetic patients. The rapid advances in the technology of data storage (20) will, however, allow greater flexibility as to the content of the stored data and, at the same time, render an estimate of future storage costs more difficult. With present use of commercial time-sharing facilities, cost estimates of the management system are as follows: Program Storage $600/year Patient Data Storage $20/year/patient CPU and Terminal Time $30/year/patient The costs of program development involve personnel, and in the case of the system reported by Bolinger et al. (1973) were approximately: Programmer-analyst 1/2 man-year Physician 1/8 man-year Nurse-clinician 1/8 man-year In addition to this, computer time for testing amounted to about $600. It would appear that in general the development of the use of the computer in the management of diabetes is relatively dependent upon similar developments in the field of health care in general. Computerized systems for the management of diabetes should be developed at a national level in parallel with development of management systems for other chronic disease states and should thus materially decrease both developmental and operational costs. A very urgent need now is some type of coordination of the developmental efforts going on at several different, presently independently, operating sites in the country. REFERENCES 1. Bartholomay, AF 1971. Some mathematical aspects of the medical diagnostic process. I. A general mathematical model. Biophysics 33:413-424. 2. Bastenie, PA, M Bonnyns, L Vanhoelst, P Neve, and M Seaquet 1971. Preclinical hypothyroidism: A risk factor for coronary heart disease. The Lancet 1:203-204. 3. Berkeley, C 1966. Development of computer applications in biomedicine. Ann NY Acad of Sciences 128:721-730. 54 Diabetes Mellitus 10. 11. 12. 13; 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. Bolinger, RE, S Price, and JL Kyner 1971. Computerized management of the outpatient diabetic. JAMA 216:11, 1779-1782. Bolinger, RE, S Price, and JL Kyner 1973. Experience with automation as an aid in the management of diabetes. Diabetes 22:480-484. Cahill, GF, JS Soeldner, GW Harris, and RO Foster 1972. Practical developments in diabetes research. Diabetes 21 (Suppl 2):703-712. Collen, MF, JL Cutler, and AB Siegelaub 1969. The reliability of self-administered medical questionnaire. Arch of Int Med 123:664-681. Dixon, RA 1970. A model of a hospital's patient medical information system. Meth Inform Med 9:88-97. Gatewood, LC, E Ackerman, JW Rosevear, GD Molnar, and TW Burns 1968. Tests of a mathematical model of blood glucose regulatory system. Computers and Biomedical Research 2:1-14. Gatewood, LC, E Ackerman, JW Rosevear, and GD Molnar 1970. Modeling blood glucose dynamics. Behavioral Science 15:72-87. Gleser, MA, and MF Collen 1972. Towards automated medical decisions. Computer and Biomedical Research 5:180-189. Goldberg, M, SB Green, ML Moss, CB March, and D Garfinkel 1973. Computer-based instruction and diagnosis of acid base disorders. JAMA 223:3, 269-275. Gottlieb, GL 1972. An approach to automated medical interviews. Computers and Biomedical Science 5:99-107. Greenes, RA, AN Pappalardo, CW Marble, and GO Barnett 1969. Design and implementation of a clinical data management system. Computers and Biomedical Research 2:469-485. Grodsky, GM, D Curry, H Landahl, and L Bennett 1969. Further studies on the dynamic aspects of insulin release in vitro with evidence for a two-compartment storage system. Acta Diabet Lat 6 (Suppl I):554. Grossman, JH, O Barnett, MT Maguire, and DB Swedlow 1971. Evaluation of computer acquired patient histories. JAMA 215:1286-1291. Hecker, R 1972. The investigation of the patient: Modern developments included automated multiphasic health screening and the use of computers in medicine. Med J Australia 2:492-496. Himbert, PJ, J Bensaid, and C Roussel 1971. Premier resultats de 1l'application aux maladies cardiovasculares d'une nouvelle conception de la medicine preventive utilisant le traitenunt automatique de l'information par ordinateur. Arch Mal du Coeur 64:538-552. Hirschfeld, WJ 1971. Forecasting and chronic illness. Bulletin of Mathematical Biophysics 33:425-437. Houston, GB 1973. Trillion bit memories. Datamation. October 52-58. Hurst, JW, and HK Walker 1972. The problem oriented record system. MEDCOM Learning Systems, New York. Kanner, IF 1969. Programmed medical history-taking with or without computer. JAMA 207:317- 321. Kadish, AH 1964. Automation control of blood sugar I. A servomechanism for glucose monitor- ing and control. Amer J Med Electronic 3:82-86. Karp, M, M Brown, and Z Laron 1972. Interpretation of glucose and insulin curves during the oral glucose tolerance test. Israel J Med Sci 8:765-766. 25. 26. 27. 28. 20. 30. 31. 32, 33. 34. 35. The Computer in the Management of Diabetes 55 Mayne, JG, W Wexel, and PN Sholtz 1968. Towards automating the medical history. Mayo Clinic Proceedings 43:1-25, Menn, SJ, GO Barnett, D Schmechel, WD Owens, and H Pontoppidan 1973. A computer program to assist in the care of acute respiratory failure. JAMA 223:308-312. Peterson, ND, JT Pearlman, BR Straatsma, and EK Bauschek 1970. Diabetic retinopathy and computer processing. Am J Ophthalmology 70:548-557. Silcock, DH, DR Hadden, and DW Neill 1972. Computer analysis of intravenous glucose tolerance tests. Diabetologia 8:301-304. Slack, W, GP Hicks and CE Reed 1966. A computer based medical history system. NEJM 274: 194-198. Sox, HC, Jr, CH Sox, and RK Tompkins 1973. The training of physicians' assistants. NEJM 288:818-824. Srinivasan, R, AH Kadish, and R Sridhar 1970. A mathematical model for the control mechanism of free fatty acid-glucose metabolism in normal humans. Comput and Biomed Res 3:146-166. Thomas, JC, and HL Dobson 1973. A functioning computer records system in a diabetic clinic. Proceedings of the 8th Congress of the International Diabetes Federation. The University Group Diabetes Program 1970. Diabetes 19 (Suppl 2):747-830. Weed, LL 1969. The problem oriented medical record. Press of the Case Western Reserve University, Cleveland, Ohio. Weed, LL 1971. Medical records, medical education, and patient care. The Press of the Case Western Reserve University, Cleveland, Ohio. oo iE 3 ie ii [5 . x | 1 Lip fhe be 12 IT ET pry RH et Bi . iy } > a x § ia inert 0 - aE Eel la Ch 4 a ETIOLOGY Chapter 6 Islet Cell Dysfunction 59 A. H. Rubenstein and D. F. Steiner Chapter 7 Infectious and Immune Mechanisms in the Etiology and/or Pathogenesis of Diabetes Mellitus 73 Bryce L. Munger rs i Po = iii EEA 3 ISLET CELL DYSFUNCTION A. H. Rubenstein and D. F. Steiner BACKGROUND Despite the contribution of more than a half century of intensive investigation, diabetes mellitus continues to be a poorly understood and highly destructive disease. As is not unusual in such a situation, there is a voluminous literature on the subject and an abundance of specu- lation about possible etiologies. Although there is convincing evidence that the disorder is genetically determined, no altered protein or gene product has been identified which might account for the predisposition of certain individuals to develop the disease or provide an accurate marker to aid in its early detection. Thus there is still room today for even the most general theories as to the origin of diabetes. A large body of evidence suggests that the inherited alteration may be confined largely, if not entirely, to the islet, or beta cell organ. Thus the tendency to diabetes is presumably expressed in the form of an abnormal protein, or as an excess or deficiency of some normal con- stituent in the beta cells of the islets of Langerhans. This alteration manifests itself at some time during the life of the predisposed individual as an impaired ability to produce suffi- cient insulin to maintain normal metabolic homeostasis in a given genotypic or environmental situation, and clinical diabetes then appears. In accord with this hypothesis, to understand diabetes we must familiarize ourselves with detailed mechanisms of differentiation, function, and regeneration of the beta cell organ. At present, little precise information is available regarding the origin and mechanisms of differentiation of the islet tissue, or the regulation of the total beta cell mass. On the other hand, considerable progress has been made in under- standing the functional activities of the beta cell and in examining these for abnormalities which might be causally related to diabetes. : As the unique function of the beta cell is the biosynthesis, storage, and secretion of insulin under the influence of various physiological stimuli, these processes will be examined in some detail in an attempt to develop an understanding of the biochemistry of this cell and lay the framework for interpreting the various hypotheses that have been advanced to explain the defect (s) in diabetes mellitus. CURRENT STATE OF KNOWLEDGE a. Insulin Biosynthesis Prior to 1967 the view was widely held that insulin synthesis in vivo was accomplished by combination of separately synthesized A and B chains. This hypothesis was supported by the initial observations of Dixon and Wardlaw (5) that small amounts of insulin could be reconstituted from mixtures of reduced insulin A and B chains in the presence of cysteine. The first direct evidence of the existence of a single chain precursor form of insulin came from studies of insulin bio- synthesis using a human islet cell tumor (32). These studies showed that tritium-labeled leucine 59 60 Diabetes Mellitus or phenylalanine was incorporated into a higher molecular weight insulin-like protein during incubation of slices from the tumor in vitro. The higher molecular weight protein was charac- terized in terms of its molecular weight, immunoreactivity with insulin antisera, and structure. Advantage was taken of the differences in distribution of phenylalanine and leucine in human insulin to show that after digestion with small amount of trypsin, the precursor gave rise to an insulin-like component containing A and B chains. It was noted that reduction of the higher molecular weight component prior to tryptic digestion did not release free A or B chains, nor did it significantly alter its molecular size as judged by gel filtration, as would have been expected for a protein consisting of a single polypeptide chain. Phenylalanine, the amino terminus of the B chain, was found to be amino terminal in the precursor as well. This observation sug- gested that the most likely structure of the precursor (which was named proinsulin) was: NH,-B chain-connecting peptide- A chain-COOH; this postulate was subsequently confirmed (2, 28). Although some further characterization of the labeled material from this tumor as well as other sources was possible, it was clear that larger amounts of the precursor form would be required for full chemical characterization. Since the precursor behaved very much like insulin in a number of respects, the possibility that it might occur as a minor component in insulin preparations was considered. Gel filtration was chosen as the initial means of purification, since this procedure has been used to separate labeled proinsulin from insulin. In fact, only 2-3 percent of the total protein in commercial crystalline insulin preparations eluted at the position of proinsulin (30). Proinsulin consists of a single polypeptide chain ranging in size from 78 (dog) to 86 (human, horse, rat) amino acid residues (Figure 1). The B chain amino acid sequence of proinsulin com- prises the amino-terminal portion of the protein and the A chain sequence comprises the carboxyl- terminal portion of the chain (2, 19). Joining the chains is a connecting segment of approximately 30 residues. The amino acid sequences a in Figare 2. There are numerous differences in their amino acids in contrast with the insulins in these species. At each end of the connecting segments two basic amino acids form connections to the amino-terminal residue of the A chain and carboxyl-terminal residue of the B chain. The remainder of the segment, aside from these residues, has been designated the C-peptide. These junctional regions, which presumably represent the sites of cleavage by an iE Oi ations in the beta cell to liberate insulin, are identical in all species of proinsulin which have been examined. Two chain intermediate forms of bovine proinsulin have been isolated in which the polypeptide chain has been cleaved either at the junction with the amino-terminus of the A chain, with loss of the lysine and arginine from positions 59 and 60, or at the junction with the carboxyl-terminus of the B chain with loss of the two arginines from positions 31 and 32 (see Figure 3). Significant progress has been made in the elucidation of the ultrastructural and biochemical organization of the process of insulin biosynthesis and secretion. Studies on a wide variety of secretory cells suggests that all such cells are organized along closely similar lines. The bio- synthesis of the secretory proteins occurs in the rough endoplasmic reticulum on membrane bound ribosomes. The newly formed secretory products are then transferred through a transitional zone of the endoplasmic reticulum, possibly in small vesicles termed microvesicles, to the tubular \ab WA oeeeis Al 82 \ [5 Gly 8 Gy). 50 “5 NEG OS ONIDINNG o.oo FIGURE 1. 1 2 3 4 5 6 7 8 5 Structure of bovine proinsulin, 9 70 A-CHAIN B-CHAIN Die CabesEeeesPRy i 0 11 12 13 14 15 NH3* - Glu - Alo - Glu- Asp- Leu - Gin- Val - Gly - Gln - Vol - Glu - Leu - Gly - Gly - Gly - NH3* - Glu - Alo - Glu - Asp - Pro - Gin - Val - Gly - Gin - Vol -Glu - Leu - Gly - Gly - Gly - ’ - Glu Alo - Glu - Asp -Pro - Glin - Vol - Gly - Glu - Vol - Glu - Leu - Gly -Gly -Gly - NH3* - Glu - Vol - Glu - Asp -Pro - Gin - Vol - Pro- Gin - Leu-Glu - Leu - Gly - Gly - Gly - NHy* - Glu - Val - Glu - Asp - Pro - Gin - Vol - Ala - Gin -Leu-Glu - Leu - Gly - Gly - Gly - © NHy > - Glu - Alo - Glu - Asn - Pro - Gin - Alo - Gly - Ala - Vol -Glu - Leu - Gly - Gly - Gly - NH3* - Glu - Vol - Glu -Gly - Pro - Gin - Vol -Gly - Alo - Leu -Glu - Leu - Ala - Gly - Gly - NHy* - Asp-Vol - Glu - 1 17 18 19 20 21 - Pro-Gly - Alo- Gly - Ser - Leu-GIn-Pro-Leu-Alo-Leu- Glu - Gly - Ser - Leu- Gin - CO - Pro-Gly - Alo- Gly - Ser- Leu-Glin-Pro-Leu-Alo-Leu-Glu-Gly- Ser -Leu-Gin -COp™ -Pro-Gly - Leu-Gly - Gly - Leu-Gin-Pro-Leu-Alo-Leu-Alo - Gly - Pro -Gln - Gin - COp™ - Pro-Glu- Alo -Gly - Asp-Leu-Gin-Thr-Leu-Alo-Leu-Glu - Vol - Alo - Arg -Gin -COp™ - Pro-Gly - Alo - Gly - Asp-Leu-Gin-Thr-Leu- Ala-Leu-Glu - Vol - -Leu-Gly - -Pro-Gly - Alo-Gly - Gly - Leu- -Pro-Gly - Glu -Gly - Gly - Leu-GIn-Pro-Leu- Ala-Leu-Glu - Gly - Ala -Leu-Gin - COp~ FIGURE 2. C-peptides. -Leu-Alo-Gly -Alo- 22 23 24 25 26 27 28 29 30 MN - Gly -Leu-Gin- Alo-Leu- Alo-Leu-Glu - Gly - Pro - Pro - Gin - COp~ - Glu - Gly - Pro - Pro - Gin - COp™ 0REenEas, Ala - Arg- Gin - COp~ Islet Cell Dysfunction 61 SEhEE0esangs J QS » ) Ere) & Gate showing sites of cleavage by trypsin. MAN MONKEY HORSE RAT I RAT I PIG Cow, LAMB DOG MAN MONKEY HORSE RAT I RAT I PIG Cow, LAMB DOG Amino acid sequences of several mammalian proinsulin These sequences do not include the basic residues at either end which link the C-peptide to the insulin chains in the proinsulin of these species. 62 Diabetes Mellitus PROINSULIN CONVERSION PROINSULIN oo (30-402) INTERMEDIATE I INTERMEDIATE IX vo] INSULIN AND C-PEPTIDE FIGURE 3. Structure of the two principal intermediate forms of bovine proinsulin. The products of the conversion of pro- insulin to insulin in the beta cell are demonstrated. elements in the periphery of the Golgi apparatus (11). The Golgi apparatus performs the function of packaging the newly formed secretory products and is also known to be the site of some biochemical transformations such as the addition of carbohydrate side chains to certain proteins. Secretory products leaving the Golgi region in newly formed granules or "condensing vacuoles,' undergo changes in morphology which indicate a continuing biochemical reorganization of their secretory contents. After 40 minutes to one hour, the newly formed proteins begin to be secreted from the cell. By means of electron microscopic radioautography a similar movement of newly labeled protein in beta cells has been observed by Howell et al. (10). Their results indicate that newly synthesized proinsulin is first transported to the Golgi region and is maximally concentrated in that region about 30 minutes after biosynthesis. At later times radioactivity is found predominantly in granules. From biosynthetic studies with isolated islets from rat pancreas as well as several human pan- creatic adenomas, it is clear that the transformation of proinsulin to insulin is a slow process that starts about 10 to 20 minutes after the beginning of biosynthesis and continues for a period Islet Cell Dysfunction 63 of hours, exhibiting a half time of about one hour (27). The initial delay can be interpreted as an indication of the transport of the newly synthesized proinsulin from the rough endoplasmic reticulum to the Golgi region of the cell. Further evidence to strengthen this supposition derives from studies on the energy requirements for the transformation of proinsulin to insulin: Addition of the potent inhibitor of mitochondrial oxidative phosphorylation, antimycin-A, strongly inhibits the transformation of proinsulin to insulin, but only when added within the first 30 minutes after biosynthesis of the proinsulin has commenced. When added at later times, antimycin has no effect on the transformation. Experiments with other inhibitors of cellular energy metabolism also indi- cate the existence of this critical early energy requirement, presumably representing a process by which the newly synthesized proinsulin is made available to the proteolytic system that will convert it to insulin (9, 29). In view of the close correlation between the time of onset of conversion, its associated initial energy dependence, and the radio-autographic data mentioned earlier indicating that the initial period after labeling is a time of transfer of the protein from the rough endoplasmic recticulum to the Golgi apparatus, we can tentatively conclude that the transformation of proinsulin to insulin begins in the Golgi apparatus. In addition, because of its long half-life, it is probable that the conversion process continues for a period of many hours after new secretory granules have been formed from the Golgi apparatus. As a consequence of the sequestration of the biosynthetic products within the relatively impermeable membranes of the secre- tion granules, the C-peptide is retained after the conversion of proinsulin and stored on an equimolar basis with the insulin. Rubenstein et al. (24) have shown that stimulation of insulin secretion is accompanied by the liberation into the circulation of equimolar amounts of the C-peptide in several animal species (Figure 4). On treatment of bovine or porcine proinsulin with trypsin, cleavages occur rapidly at the sites of attachment of the C-peptide to the chains of insulin and subsequently at position B-29 (lysine), liberating fully active dealanated insulin and the C-peptide with or without lysine remaining at the carboxyl-terminus. In order for an enzyme having the chemical specificity char- acteristic of trypsin to liberate native insulin from proinsulin in vivo, a second enzymatic activity similar to that of carboxypeptidase-B would be required. Such an enzyme could remove carboxyl-terminal basis residues which would originate through the action of a trypsin-like enzyme. Recently it has been shown that trypsin and carboxypeptidase-B together can carry out the cleavages found in the intermediate components in vitro and thus give rise to intact insulin and C-peptide (12). Nevertheless, it has been difficult to demonstrate conclusively that similar enzymes are involved in the conversion process in vivo (13). Although it is well established that glucose is an important stimulus to insulin synthesis, the mechanism by which it exerts this effect is not yet understood (31). It is of particular interest that the stimulating effect of glucose is strongly selective for insulin biosynthesis: the synthesis of other cellular proteins being enhanced to a far smaller degree (22). Moreover, the glucose stimulus is not dependent on new RNA synthesis, but rather appears to be due to a selective translational enhancement. Thus actinomycin D initially does not inhibit the stimulation of bio- synthesis due to glucose, although it does appear to inhibit a subsequent phase of further enhance- ment of the biosynthetic rate, that may depend in some way on additional RNA synthesis. Furthermore, the stimulatory effect of glucose on biosynthesis is inhibited by mannoheptulose (15), suggesting a requirement for glucose metabolism in the generation of the response. 64 Diabetes Mellitus BETA GRANULE FORMATION R.E.R. AMINO ACIDS* TRANSFER RNA ATP, GTP, Mg++ 10-20 Min. Enzymes PROINSULIN* (S-S Bond ANTIMVCIN BLOCKS formation) So o ~ 1 M. o ore TRANSFER STEP u MY (energy dependent) . V. 9 eo iN* 5% {-TRAlgres STEP 2 20 Min. di (O+-EARLY GRANULES PROINS GOLGI Zinct* MEM ONG or PROGRESSIVE BOUND Arg++ | an ong) Ln CONVERSION (tl =~1 hour) -1 i A MAINLY INSULIN ® 30-120 Min (Crystalold) = / + PA O \ MEMBRANE C-PEPTIDE ® atin os \ ® MATURE ? ° HOURS-DAYS erin (Space) GRANULES — PLASMA TRANSFER STEP 3 MEMBRANE ° (Energy depend nt Ca++ dependent) EMIOCYTOSIS SECRETED PRODUCTS (EXOCYTOSIS) ir a TT INSULIN ] 94% C-PEPTIDE PROINSULIN ~6% INTERMEDIATES ++ OTHERS ? FIGURE 4. Diagrammatic representation of the insulin bio- synthetic and secretory mechanism of the beta cell (R.E.R. = rough endoplasmic reticulum: M.V. = microvesicles). b. Insulin secretion At present most investigators believe that the defect in diabetes will be found in the intricate mechanism concerned with the secretion of insulin. The system has an afferent component which is involved in monitoring the ambient glucose concentration and an efferent component which is concerned with the secretion of insulin from the beta cell. Although a number of critical steps in these pro- cesses are still unknown, a great deal of progress has recently been made in unravelling its molecu- lar basis. There is now a great deal of evidence that the major path way of insulin secretion is by way of exocytosis (14). In this process, the granule membrane fuses with the plasma membrane. The insulin granules are subsequently extruded into the extracellular space where they undergo dissolution. When insulin release is markedly stimulated, an increased number of cytoplasmic projections, or microvilli, can be seen on the beta cell surface, presumably as a result of the addition to it of many granule membranes. Orci (21) has demonstrated that these microvilli are subsequently reincor- porated into the cytoplasm as vesicles. During the past 6 years, increasing application of improved electron microscopic techniques to islet morphology have led to an appreciation of the intracellular structures involved in the secretory Islet Cell Dysfunction 65 process. Lacy et al. (14) first suggested that a microtubular-microfilamentous system was involved in the movement of insulin granules towards the plasma membrane of the beta cell. These organelles are composed of actin-like material and are found in close association with mature insulin granules (Figure 5). It is believed that the microtubules may direct the granules towards the cell surface while the microfilamentous web might act as a barrier which controls the access of granules to the cell membrane (16). The use of agents which selectively disrupt these structures has lent support to their involvement in the secretory process. Thus colchicine and vincristine, which interact with microtubular protein and lead to its disappearance or precipitation as crystalline-like material, inhibit insulin secretion in response to glucose or glucose and theophylline. Deu- terium oxide, a stabilizer of microtubules, reversibly inhibits secretion stimulated by glucose, leucine, and tolbutamide. Exposure of islets to cytochalasin B, on the other hand, enhances insulin release, presumably by causing a spatial reorganization of the microfilamentous material and margination of granules. A B Cc D E FIGURE 5. Schematic representation of the relationship between secre- tory granules, microtubules (mt), and the microfilamentous cell web. In the unstimulated beta cell, the secretory granules are kept away from the A plasma membrane by the cell web (A). Under stimulation, granules are transported along the microtubules, and the web might participate in their access to the cell membrane (B). After fusion between the granule membrane and the plasma membrane, an emiocytotic aperture occurs (C), and the gran- ule core is extruded into the extracellular space (D). The incorporation of the membranous sacs encasing the granules into the plasma membrane ap- parently results in the formation of microvillous processes (mp;E). (From Malaisse 1973.) Reprinted with premission from Springer-Verlag (Diabetologia 9:167, 1973). The concept that insulin release is triggered by activation of the microtubular-microfila- mentous system has been strengthened by the demonstration that extracellular calcium is required for this process (16). Thus exposure of islets to glucose in the presence of 45-calcium results in a net accumulation of this ion. If the islets are prelabeled with wadioactive calcium, en- hancement of secretion is associated with an immediate reduction in calcium efflux. Thereafter, a marked increase in calcium extrusion, associated with insulin release, occurs. It is postu- lated that the accumulation of calcium from the extracellular fluid, or the redistribution of the jon from intracellular storage sites within beta cells is a necessary prerequisite for activating . 66 Diabetes Mellitus insulin secretion. The movement of calcium appears to be coupled to that of another ion, namely sodium (Figure 6): Studies in other systems have shown that calcium uptake is inhibited by high extracellular sodium concentrations, but is enhanced when the intracellular sodium level is raised. It seems probable that two intracellular sodium ions are exchanged for one extracellular calcium ion. The inhibiting effect of diphenylhydantoin, a drug known to reduce intracellular sodium concentrations in brain and muscle, on insulin release, as well as the stimulating action of ouabain, which has the opposite effect, suggest that the sodium dependent calcium uptake also exists in beta cells (23). Consideration of the above findings suggested that depolarization of the beta cell membrane might be an important early event in insulin release. The elegant experi- ments of Dean and Matthews (4) using ultra microelectrodes to record transmembrane potentials in mouse beta cells under basal and stimulated conditions has confirmed and extended these hypotheses. METABOLIC REGULATION glucose leucine J | } { SODIUM-DEPENDENT UNKMOWN METABOLISM+TRANSPOR SIGNAL EXTRACELLULAR o£ CYTOSOLIC pee INSULIN RELEASE CALCIUM = ® @ [cAMP |=—{ATP CALCIUM =P7©) 7 TN i+ Mg Nat VACUOLAR CALCIUM adenylcyclase epinephrine enteroglucagon HORMONAL MODULATION FIGURE 6. An integrated model for the multifactorial regulation of insulin secretion by metabolic and hormonal agents. On the left side, the T-shapes bars represent the hypothetical sites of competition or f inhibition of calcium transport across the cell membrane. (From Malaisse 1973). Reprinted with permission from Springer Verlag (Diabetologia 9:167, 1973). Beta cell cyclic 3', 5'-AMP appears to play a role in regulating insulin release (Figure 6), similarly to its effects in modulating hormone secretion in other endocrine glands. Agents which increase its concentration either by stimulating adenyl cyclase (such as glucagon, gastrointes- tinal hormones or beta adrenergic stimulators) or by inhibiting cyclic 3', 5'-AMP phosphodies- terase (such as caffeine and theophylline) lead to an enhancement of insulin release (17). The precise mechanism whereby cyclic 3', 5'-AMP affects insulin secretion is controversial. Three possibilities which have received attention involve (1) activation of an enzyme which may phos- phorylate the microtubular protein (2) alteration of the intracellular distribution of calcium so as to increase the calcium concentration in the vicinity of the microtubules (3) modulation of a rate limiting step in glucose metabolism. However, it should be realized, that whichever mechanism turns out to be correct, the available evidence does not support a direct role for cyclic AMP in glucose induced insulin secretion. This conclusion is based on the inability of dibutyryl- cyclic-AMP to enhance insulin release in the absence of glucose and the finding that glucose Islet Cell Dysfunction 67 induced insulin secretion is not associated with detectable changes in islet cyclic-AMP levels. It has been known for many years that the level of glucose in arterial blood reaching the pan- creas is the most important factor regulating insulin secretion. However, the mechanism whereby glucose exerts its effect is still uncertain. Two main theories have been considered (18). The first involves the concept of glucose combining with a specific membrane receptor, presumably a protein, which activates release, either directly or by stimulating a messenger molecule, while the second allows for the metabolism of glucose to a substrate which would be the trigger for insulin release. Part of the evidence favoring the first hypothesis has been derived from analysis of the sigmoid relationship of glucose concentration to the rate of insulin release; the rapidity of insulin release after exposure to glucose; measurements of intracellular metabolite levels (which may show no changes at times when insulin release is stimulated), as well as alterations in the sensitivity of the insulin release mechanism to glucose after fasting, and the inhibition of the restoring effect of refeeding by actinomycin D (8). In addition, the finding that certain non- metabolisable amino acids stimulate insulin release has added weight to this concept (6), as has the recent demonstration that the two anomers of D-glucose, which are metabolized similarly, have different potencies in stimulating insulin release. The major arguments for the metabolism hypo- thesis’ are based on experiments showing that mannoheptulose, which inhibits the phosphorylation of glucose, blocks its effect on insulin secretion. In addition, non-metabolizable sugars, or those that can only be phosphorylated, do not release insulin (34). It is obvious that further informa- tion will be required to resolve these theories. c. Beta cell dysfunction in diabetic patients With the development of an immunoassay for insulin, it has become increasingly accepted that diabetes usually results from some degree of secretory failure of the pancreatic beta cells. In juvenile diabetics the extent of this failure is severe and is reflected in gross destruction of islet tissue. In adults diabetics, secretory failure is less pronounced, but when patients are carefully classified so that variables, such as obesity, are controlled, some degree of impairment of insulin secretion is almost always observed. In the glucose tolerance test there is both a quantitative decrease in total insulin secretion, as well as a sluggish early response with a tendency for the peak level to occur later than normal. Whether the time course of the insulin response provides a significant clue to the nature of this defect is a controversial question, which has received much attention (1), especially in terms of the concept of two separate phases of insulin secretion, an early rapid burst followed by a later, more prolonged phase (3). On the other hand, the initial delay in secretion generally cor- relates well with the tendency in these tests for the blood sugar level to rise to higher levels and to peak at later times. It has been claimed on the basis of these results that islet cells of diabetics may have an inherent or acquired alteration in their sensitivity, or ability to respond “to a glucose stimulus. This concept is also supported by the observations that other stimuli to insulin secretion, such as tolbutamide and glucagon, elicit normal secretory responses in mild diabetics, at a time when the response to glucose is already impaired (25). Studies of the pathologic changes in the pancreatic islets in adult onset diabetes support the concept of a primary failure of islet responsiveness. Gepts (7) has pointed out that there is almost invariably a reduction in total islet tissue mass in the diabetic pancreas, amounting to approximately 50 percent in many cases. Moreover, there is a reduction of insulin stores 68 Diabetes Mellitus in the surviving pancreatic beta cells of these individuals as evidenced by partial degranulation of the islet tissue and by a decrease in the total insulin that can be extracted from the pancreas. These pathological data suggest that the defect may involve the production of insulin and the regeneration of islet cells, as well as the secretion of the hormone. It is interesting to hypo- thesize that these defects may have a common and interdependent origin, in the context of an altered glucose receptor mechanism in the diabetic's beta cells. Such an alteration could possibly contribute to the impaired renewal of islet tissue through failure to adequately stimulate cell division, for there is a suggestion that hyperglycemia may play a role in stimulating mitotic activity in the islets. In addition, this failure to respond normally to glucose could lead to a decrease in insulin biosynthesis and storage, because the glucose concentration is known to be a potent stimulus for this process. Finally, the failure of an adequate mechanism to monitor the extracellular glucose concentration would, of course, result in impairment of insulin secretion. Although it is possible that the beta cell defect may affect only the efferent component of the insulin release mechanism, one would anticipate that the cellular stores of insulin would not only be preserved, but might be even greater than normal under these conditions. One might anticipate also that islet cell regenerative activity would lead to islet cell hyperplasia and cell proliferation. This situation has been found in the diabetic spiny mouse (26), where there is a great increase in total islet tissue mass and the beta cells contain numerous secretion granules and large quantities of insulin. Recent evidence has indicated that these animals have an intrinsic defect in their secretory mechanism, probably involving the microtubular-microfila- mentous system. While the concept of a genetically determined intrinsic defect in the beta cells of most diabetics is an attractive working hypothesis, other factors, of largely environmental origin, merit further consideration. Recent studies of the inheritance pattern of diabetics and its incidence, particularly in identical twins (33) strongly suggest that other causes for diabetes may exist and account for a significant fraction of the total number of patients. Studies of pancreatic pathology, particularly in juvenile diabetics, indicate the occurrence of a complex destructive lesion in the islets of Langerhans that may be due to extrinsic causes acting in a genetically favorable situation. Among such causes are two of particular interest and concern: autoimmunity and viral infection. These subjects will be dealt with in other chapters in this monograph. Although there is little evidence at present to support the idea of a defect of the receptor for insulin in diabetes, recent studies with lymphocytes have indicated that abnormalities in membrane-binding of insulin may occur under certain conditions (20). Elucidation of the important structural features of insulin necessary for its biological action, as well as the molecular events associated with its biological effects, are both of obvious importance to the complete under- standing and successful treatment of diabetes. APPRAISAL OF INFORMATION WHICH NEEDS TO BE ACQUIRED THROUGH RESEARCH The information in this chapter is concerned with the fundamental defect (s) which give rise to the disease diabetes mellitus. Although there is considerable evidence that abnormalities in beta cell structure or function may be etiologically involved in the expression of diabetes, there is not, as yet, absolute certainty on this point. Clearly, therefore, any experimental attack on Islet Cell Dysfunction 69 the problem of human diabetes must be carried out on a very broad front. It is important to appreciate that the development of more basic knowledge concerning the regulatory mechanisms of the beta cells, their growth, replication, and biosynthetic processes, as well as their secretory mechanisms, are of utmost importance to the ultimate elucidation of the diabetic defect. Moreover, it is impossible to predict how new information derived from other areas of biological investiga- tion may change our concepts regarding the causes of diabetes. Nevertheless, it seems safe to suggest that as our information of the normal structure and function of beta cells expands, the possibility of pin-pointing abnormalities in diabetes becomes more likely. The implication in the above comments is that it is possible that no one single defect under- lies every case of diabetes. Many diverse abnormalities in beta cell integrity may result in a clinical syndrome characterized by carbohydrate intolerance. Moreover, one must also bear in mind that many insults, both genetic and environmental, may affect a specific process. Until all these ramifications are dissected out, it would be as well to pursue both basic and clinical research into many areas of beta cell function. Among the many approaches that may be profitable, one might mention experiments designed: a. To elucidate the natural history of beta cells. Their mitotic potential and factors which may influence or initiate this process need to be determined. Information about the length of time beta cells survive should be sought. The capacity to maintain islets and the beta cells in organ and monolayer culture respectively will undoubtedly become one of the most important methods to pursue these questions. b. To determine the progenetor cells of the pancreatic islets. Do beta cells arise from proliferating pancreatic ductules which are derived from the original doudenal diverti- cula? Do diabetic patients have a normal complement of beta cells before their disease manifests? These questions will require quantitative morphological and functional studies on beta cells at various stages of gestation in both animals and humans. c. To determine the factors regulating insulin biosynthesis and the mechanism of their effect. Isolation and characterization of the converting enzymes which transform pro- insulin to insulin is also necessary for the complete understanding of insulin formation. The existence of a precursor form of insulin has provided an enzyme mediated step in biosynthesis where inherited defects could occur. Although initial studies in diabetic patients have failed to support the hypothesis of a defect in conversion of proinsulin to insulin, further investigations on a more sophisticated level are required. d. To study the amino acid sequence of proinsulin and insulin in diabetics. The genetic defect of diabetes may express itself through point mutations in the structural gene for proinsulin. Amino acid substitutions in the connecting peptide region of proinsulin could distort the folding of the peptide chain so as to reduce the effectiveness of disulphide bond formation, while alterations in the sequence in the regions of proinsulin cleavage could also lead to abnormal products. Of course, mutations within the insulin chains themselves could result in biologically ineffective insulin molecules. Refine- ments in techniques for extracting these proteins from single human pancreata, and the a- bility to determine their amino acid sequences using very small quantities of material will be needed to accomplish these aims. e. From knowledge of the amino acid sequences of proinsulins from several mammalian species, 70 Diabetes Mellitus certain predictions can be made regarding the structure of the gene(s) in the nuclear DNA that encode the prohormone. Recent advances in techniques for the isolation of genes make it technically feasible to now undertake the isolation of the genes for proinsulin. To do this will first require the isolation of the proinsulin messenger RNA from the polyribosomes of beta cells actively engaged in the synthesis of insulin. This mRNA can then be used to transcribe copies of one of the DNA strands of the pro- insulin gene for use as a probe in identifying the proinsulin gene in the chromosomal DNA. The ultimate availability of 'proinsulin-DNA" will also enable measurements to be made of the amount of proinsulin mRNA per beta cell, and it should then be possible to learn more about how insulin production is regulated by glucose, cAMP, and other factors at the levels of genetic transcription and of translation into protein. f. To resolve the issue of whether a structural gluco-receptor is present in beta cells, and to determine its structure if it is found. This problem will require extensive studies on the morphology and chemistry of beta cell membranes and the correlation of these results with its biochemical properties. A long-term goal would be directed towards determining whether a defect in the ''gluco-receptor' may be present in diabetic patients. This would require new and innovative clinical and basic research initiatives. g. To determine whether insulin receptors are abnormal in diabetic patients. HOW THIS INFORMATION WOULD LEAD TO IMPROVEMENT IN PREVENTIVE MEDICINE It is obvious that a necessary first step in the train of events leading to the prevention of diabetes mellitus is an understanding of the etiology of this disorder. For this reason, it is at present impossible to predict with certainty what these approaches may be. However, experience gained from other diseases has suggested that the most effective preventative measures will require this information. While it is true that this is a long-term endeavor, it would be a mistake to sacrifice support for these basic projects for shortterm, but less fundamental management of the disease. COST I do not know of any way to accurately estimate the cost of the experimental plans that are needed. Perhaps one could arrive at a reasonable figure by obtaining information from the National Institutes of Health regarding all approved diabetes-related research grant applications dealing with this aspect of the disease. The total budgets of these grants, both those that have been funded and those that have not been funded because of lack of funds, may indicate a reasonable dollar estimate. REFERENCES 1s Cerasi, E, and R Luft 1972. Pathogenesis of genetic diabetes mellitus: further development of a hypothesis. Mount Sinai J Med 40:334-349, 2. Chance, RE, RM Ellis, and WW Bromer 1968. Porcine proinsulin: characterization and amino acid sequence. Science 161:165-167. 3. Curry, DL, LL Bennett, and GM Grodsky 1968. Dynamics of insulin secretion by the perfused rat pancreas. Endocrinology 83:572-584, 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22, 23. Islet Cell Dysfunction 71 Dean, PM, and EK Mathews 1970. Glucose-induced electrical activity in pancreatic islet cells. J Physiol (London) 210:255-264. Dixon, GH, and AC Wardlaw 1960. Regeneration of insulin activity from the separated and in- active A and B chains. Nature 188:721-724. Fajans, SS, and JC Floyd 1972. Stimulation of islet secretion by nutrients and by gastro- intestinal hormones released during digestion: in Handbook of Physiology, Vol. I. Endocrine Pancreas (Eds., DF Steiner and N Freinkel). Baltimore, Williams and Wilkins, pp 473-493. Gepts, W 1972. Pathology of islet tissue in human diabetes: in Handbook of Physiology, Vol. I. Endocrine Pancreas (Ed., DF Steiner and N Freinkel). Baltimore, Williams and Wilkins, pp 289-303. Grey, NJ, S Goldring, and DM Kipnis 1970. The effect of fasting diet and Actinomycin D on insulin secretion in the rat. J Clin Invest 49:881-889. Howell, SL 1972. Role of ATP in the intracellular translocation of proinsulin and insulin in the rat pancreatic B cell. Nature, New Biology 235:85-86. Howell, SL, M Kostianovsky, and PE Lacy 1969. Beta granule formation in isolated islets of Langerhans: A study of electron microscopic radioautography. J Cell Biol 42:695-705. Jamieson, JD, and GE Palade 1971. Condensing vacuole conversion and zymogen granule dis- charge in pancreatic exocrine cells: Metabolic studies. J Cell Biol 48:503-522. Kemmler, W, JD Peterson, and DF Steiner 1971. Studies on the conversion of proinsulin to insulin: I. Conversion in vitro with trypsin and carboxypeptidase B. J Biol Chem 246:6786-6791. Kemmler, W, DF Steiner, and J Borg 1973. Studies on the conversion of proinsulin to insulin. III. Studies in vitro with a crude granule fraction isolated from islets of Langerhans. J Biol Chem 248:4544-4551. Lacy, PE, SL Howell, DA Young, and CJ Fink 1968. New hypothesis of insulin secretion. Nature 219:1177-1179. Lin, BJ, and RC Haist 1969. Insulin biosynthesis: Effects of carbohydrates and related compounds. Canad J Phys Pharmacol 47:791-801. Malaisse, WJ 1973. Insulin secretion: Multifactorial regulation for a single process of release. Diabetologia 9:167-173. Malaisse, WJ, F Malaisse-Lagae, and D Mayhew 1967. A possible role for the adenylcyclase system in insulin secretion. J Clin Invest 46:1724-1734. Matschinsky, FM, R Landgraf, J Ellerman, and J Kotler-Brajtburg 1972. Glucoreceptor mecha- nisms in islets of Langerhans. Diabetes 21 (Suppl. 2):555-569. Nolan, C, E Margoliash, JD Peterson, DF Steiner 1971. The structure of bovine proinsulin. J Biol Chem 246:2780-2796. Olefsky, JM, and GM Reaven 1974, Decreased insulin binding to lymphocytes from diabetic patients. J Clin Invest 54:1323-1328. Orci, L 1974. A portrait of the pancreatic beta cell. Diabetologia 10:163-188. Permutt, MA, and DM Kipnis 1972. Insulin biosynthesis. I. On the mechanism of glucose stimulation. J Biol Chem 247:1194-1199. Randall, PJ, and CN Hales 1972. Insulin release mechanisms. In Handbook of Physiology, Vol. I. Endocrine Pancreas (Eds., DF Steiner and N Freinkel). Baltimore, Williams and Wilkins, pp 219-235. 72 24. 25. 26. 27. 28. 29, 30. 3. 32. 33. 34. Diabetes Mellitus Rubenstein, AH, JL Clark, F Melani, and DF Steiner 1969. Secretion of proinsulin C-peptide by pancreatic B cell and its circulation in blood. Nature 224:868-869. Simpson, RG, A Benedetti, GM Gradsky, JH Karam, and PH Forsham 1968. Early phase of insulin release. Diabetes 17:682-692. Stauffacher, W, L Orci, M Amherdt, IM Burr, L Balant, ER Froesch, AE Renold 1970. Metabolic state, pancreatic insulin content and B-cell morphology of normoglycemic spiny mice (Acomys cahirinus): Indications for an impairment of insulin secretion. Diabetologia 6:330-342. Steiner, DF 1967. Evidence for a precursor in the biosynthesis of insulin. Trans NY Acad Sci 30:60-68. Steiner, DF, JL Clark, C Nolan, AH Rubenstein, E Margoliash, B Aten, and PE Oyer 1969. In recent Progress in Hormone Research (Ed., EB Astwood), New York, Academic Press, pp 207-282, Steiner, DF, JL Clark, C Nolan, AH Rubenstein, E Margoliash, F Melani, and PE Oyer 1970. The biosynthesis of insulin and some speculations regarding the pathogenesis of human diabetes. In the Pathogenesis of Diabetes Mellitus. Proceedings of the Thirteenth Nobel Symposium (Ed., E Cerasi and R Luft). Stockholm. Almquist and Wiksell, pp. 123-132. Steiner, DF, O Hallund, AH Rubenstein, S Cho, and C Bayliss 1968. Isolation and properties of proinsulin, intermediate forms, and other minor components from crystalline bovine insulin. Diabetes 17:725-736. Steiner, DF, W Kemmler, JL Clark, PE Oyer, and AH Rubenstein 1972. The biosynthesis of in- sulin, in Handbook of Physiology, Vol I. Endocrine Pancreas (Eds., DF Steiner and N Freinkel). Baltimore, Williams and Wilkins, pp. 175-198. Steiner, DF, and PE Oyer 1967. The biosynthesis of insulin in and a probable precursor of insulin by a human islet cell adenoma. Proc Nat Acad Sci 57:473-480. Tattersall, RB, and DA Pyke 1972. Diabetes in identical twins. Lancet 2:1120-1125. Taylor, KW 1972. The biosynthesis and secretion of insulin. Clinics in Endocrinology and Metabolism 1:601-622, NFECTIOUS AND IMMUNE MECHANISMS IN THE ETIOLOGY ND/OR PATHOGENESIS OF DIABETES MELLITUS! I A Bryce L. Munger Diabetes mellitus in man and animals has been the subject of innumerable books, monographs, and scholarly works. Yet, in the middle of the twentieth century, we do not know in what way the B cellin the islets of Langerhans are abnormal, or if indeed they are abnormal (77). This is an amazing revelation in an era of subcellular and molecular pathology. The disease entity, diabetes mellitus, has survived the era of tissue pathology prior to the 1850's and the era of cellular pathology from the time of Virchow (circa 1856) to the mid-twentieth century, and the recent era of biochemical and molecular pathology without a satisfactory explanation as to the basic etiology and/or pathogenesis of the disease. This admission of basic lack of knowledge may be hard for scholars to accept, but it is a fact. Several factors have contributed to this present state of knowledge, or rather the lack of it. Gepts (22, 23) has provided the most recent summary of the pathologic changes in the pancreas in acute juvenile diabetes. He collected 22 specimens of acute juvenile diabetes from various hospitals, and as one might expect, the quality of preservation varied considerably. Despite this limitation, Gepts' study is still one of few thorough light microscopic studies on the acute cytopathology of juvenile diabetes in man reported in the recent literature. This study documented the degranulation and vacuolation of B cell cytoplasm as well as a constant lymph- ocytic infiltrate and scarring in the vascular stoma of the islets. This study has strengthened the suggestion that infectious and/or immune mechanisms might be involved in the etiology and/or pathogenesis of diabetes in man, baséd in part on the omnipresent mild inflammatory reaction in the islets as reported by Gepts. We shall return to this point subsequently. In contrast to a paucity of studies concerning the acute cytopathology of the islets in human diabetes, our understanding of the pathogenesis of glomerular lesions has been facilitated by electron microscopic study of renal biopsies from living patients. The same can be said for various diseases of liver, lung, intestinal tract, skin and muscle--all amenable to biopsy and study by modern methods of cytochemistry, biochemistry, and especially electron microscopy. Electron microscopic study of the human endocrine pancreas has been based on isolated cases of surgical biopsies in cases of tumor (56). We dare not biopsy the pancreas in an acute case of juvenile diabetes, e.g., a patient age 14, and perform these kinds of studies. We have thus been unable to study acute human diabetes mellitus and have instead searched for adequate animal models, including the use of chemical “poisons, such as alloxan, which kill B cells in the pancreatic islets but does not result in the vascular, neural, optic, and renal problems en- countered in human diabetes. IThis study was supported in part by USPHS Research Grant No. AM11407 and a grant from the John A. Hartford Foundation. 73 74 Diabetes Mellitus In the ensuing discussion, juvenile diabetes will be emphasized, since the pathology of the islets is more pronounced, and the possibility of infectious and/or immune mechanisms in the pathogenesis of the disease more plausible, as contrasted with typical maturity onset diabetes. The classification of types of diabetes (i.e., clinical, asymptomatic, latent, and potential) will follow that of Cerasi and Luft (9). The above comments clearly indicate that our understanding of human islet pathology is severely restricted. This restriction also limits our understanding of the pathogenesis of the disease, especially as it relates to the islets. These limitations in terms of understanding of pathogenesis in terms of an organ are further complicated by a lack of understanding of metabolic and/or molecular changes in islet B cells as it relates to juvenile diabetes mellitus. These restrictions have led scholars to search for appropriate animal models of the disease. Our knowledge of animal models has been summarized in two Brook Lodge Workshops on Spontaneous Diabetes in Laboratory Animals (63, 64). The models to date are all clearly genetic or dietary in etiology. This brings us to another problem. Human diabetes mellitus has a genetic component, but it is definitely not solely a genetic disease, as is cystic fibrosis of the pancreas or sickle cell anemia. If a genetic etiology is to be used, partial expression, partial penetrance, multiple alleles, and other "weasel terms' must be invoked to explain the disease. So we now couple inadequate explanation of cellular pathogenesis with multiple forms and compound the issue by involing only a ''genetic component.' Since a genetic etiology is not adequate, the need clearly exists to examine other models for the etiology, as well as the pathogenesis of the disease. These terms will be used in the ensuing discussion as follows: etiology is a specific causative agent, e.g., polio virus causes poliomyelitis, a gene abnormality causes sickle cell disease, pneumococcus causes pneumococcal pneumonia, etc. The pathogenesis of a disease is its cellular and tissue life history or progression. Our concepts of the pathogenesis of tubercu- losis as expressed by Arnold Rich (71) are an adequate example of this concept. One major area of interest other than genetic as to the pathogenesis of diabetes mellitus invokes infectious agents or immune mechanisms in the etiology and pathogenesis of the disease (9). One factor missing from this approach has been the absence of an animal model in which the pathogenesis of the disease resembles that seen in man. An animal model of viral (infectious) etiology with a genetic component has recently been described by Craighead (12, 13). A second animal model of contagious etiology also possessing a genetic component has recently been studied in our laboratory (59). The ensuing discussion will first present human and animal models of infectious and immune mechanisms and close with a brief discussion of our guinea pig model of contagious diabetes mellitus and its possible implications for future study. Where possible, potential fertile research areas will be pointed out. INFECTIONS AND HUMAN DIABETES Most studies or reports of the association of diabetes mellitus with various infections are temporal associations of two clinical events. This subject has been recently reviewed by Levy and Notkins (48). A prime example is mumps and diabetes. From the early description of Harris (28), repeated examples of diabetes following mumps can be found. These isolated reports on the development of diabetes following mumps usually consist of two or three cases added to a small literature search (26,31,51,52 »67). Two very interesting cases, described by Infectious and Immune Mechanisms 75 King (35) and by Messaritakis et al. (53), report the development of diabetes in two siblings each following mumps. As evident from this literature, interest has only been sporadic with a couple dozen cases reported. The existence of this association in sibs is perhaps another matter. Certainly the criticism of Levy and Notkins (48), that neither mumps nor diabetes had been proven in those cases prior to 1940, would not hold in the case of the two siblings described by Messaritakis et al. (53). Certainly the pathogenesis of mumps-associated diabetes is a subject deserving careful attention. An association of mumps and diabetes is possible since mumps virus can involve the pancreas. The pathology of these accumulated cases has yet to be described. Thus, a few cases have been described wherein mumps and diabetes are temporally related. As an etiologic agent, mumps virus can only be suspect. These comments say nothing as to the pathogenesis of changes in the pancreatic islets or other organs. A more recent addition to the list of potential suspects are the Coxsackie viruses, specif- ically type B4. Gamble and coworkers (21, 20) have found positive correlation in seasonal incidence, as well as elevated titers of B4 antibodies, in recently acquired (less than three months' duration) diabetes. A seasonal variation has been noted since the pioneering study of Adams (1) who agrees basically with Gamble and Taylor (21). Furthermore, Coxsackie virus has been isolated from human pancreas in five cases reported by Fechner, Smith, and Middlekamp (17). To complicate matters thoroughly, Hadden et al. (27) failed to find elevated B4 titers in 58 newly diagnosed diabetics; 34 required insulin. An animal model involving Coxsackie virus will be discussed shortly. Other agents associated with diabetes are much more tenuous in their relationship as a possible causative or etiologic role. White (78), John (33), Grishaw et al. (24), and more re- cently Brown (6) have all argued for a role of infection in the onset of overt clinical dia- betes. Brown has been most persuasive in his arguments as to an infectious origin for juvenile diabetes mellitus. The list of isolated agents associated with diabetes or pancreatitis in- cludes hepatitis, poliomyelitis, influenza, tick-borne encephalitis, rubella, and cytomegalic inclusion disease (reviewed by Levy and Notkins, 48). A relationship between brucellosis and diabetes has been reported by Leon and Aguirre (46) and Harris (29); however, firm evidence is lacking. Harris (29, and personal communication) is still convinced he has seen numerous examples in his practice, and I personally have also encountered a unique case with a relat- ionship between chronic brucellosis and diabetes. The case for maternal rubella associated with the onset diabetes is also puzzling in that the evidence is recent and adequate (18,19). Forrest, Menser, and Burgess encountered two cases in their own clinical experience follow- ing former rubella cases. They sent letters to a local newspaper in Sydney, Australia, as well as a medical journal, seeking additional cases. Their search located three additional cases. The authors state that the association of diabetes and congenital rubella could be mere chance, as all five cases had a family history of diabetes. Subsequently Forrest, Menser, and Burgess (18) studied 50 young adults with congenital rubella by means of a standard two-hour glucose tolerance test in 44 of the original group of 50. Five cases of overt diabetes were detected (11 percent) and four cases (9 percent) of asymptomatic diabetes (slightly abncrmal glucose tolerance tests with delayed insulin elevation). These statistics are difficult to argue 76 Diabetes Mellitus against. A total of 20 percent diabetics in a group of 44 cases of congenital rubella is astounding, and in the words of Forrest, Menser, and Burgess (18) ''strongly suggests a causative relationship between the two conditions." We thus have questionable suspects as etiologic factors, mumps, Coxsackie viruses, and congenital rubella, as well as several more distant possibilities. If we consider the long list of associated events (upper respiratory infections in general plus numerous specific disease entities) described as related to the onset of diabetes, the situation can be considered confus- ing at best. While any of these could be an etiologic factor, what is the pathogenesis? Cer- tainly we have no evidence clearly pinning down a single infectious agent associated with the onset of human diabetes mellitus. In terms of needed research, more clinical data, including antibody titers and detailed histories, are badly needed on newly diagnosed acute diabetics. The results of Forrest, Menser, and Burgess (18) could never have been predicted from any single personal clinical experience. By looking carefully at 50 cases of congenital rubella, a 20 per- cent incidence of diabetes (overt and asymptomatic) was uncovered. Studies of this sort, i.e., looking carefully at groups of patients with mumps, Coxsackie infections in detail with glucose lerance tests, as well as plasma insulin, need to be done. Specific suggestions as to how we ought to be handling newly diagnosed diabetics will be covered later. INFECTIOUS AGENTS AND ANIMAL MODELS The pathology and pathogenesis of viral-induced pancreatic lesions date to the studies of Robertson (72) on pleurodynia virus-induced pancreatitis in mice. Due to the fact that Coxsackie viruses could be passed through adult mice and result in pancreatic lesions (66, 14), Burch et al. (7) studied the pancreatic lesions resulting from type B4 viruses (the strain reported to have elevated antibody titers in human acute diabetes) (20). Necrosis of acinar tissue and inflammation around islets did result, but not diabetes. A more specific islet lesion has been described by Craighead and co-workers (11,12,1:3), This model involves a strain of the encephalomyocarditis virus that produces selective necrosis of the islets. The lesion is indeed inflammatory involving necrosis of B cells and release of immunoreactive insulin into the circulation. In a small number of animals (5 to 10 percent), a chronic diabetic condition persisted; and in some of these animals, renal glomerular lesions were described. In addition, genetic factors seemed to be involved since the DBA strain of mice are susceptible, whereas only infrequently are CH mice affected. Unfortunately, the lesions in the pancreas do not mimic the pathology of human diabetes mellitus (22, 23, 77). The validity of this animal model must await further studies on both the pancreatic islets and other systems (renal, vascular, occular) usually involved in diabetes. These studies by Craighead et al. provided some logic for an editorial in Lancet (16) which discussed the fact that Burch's work might be relevant to Craighead's, since their viruses were all from the picornavirus group (RNA viruses) (poliomyelitis, ECHO, Coxsackie, encephalomyocard- itis, and foot-and-mouth disease). What is even more remarkable is that in a spontaneous out- break of foot-and-mouth disease in Italy, some cattle developed diabetes (4). This group sub- sequently produced the disease experimentally in other cattle (5). The lesions consisted of inflammatory cell infiltrates in the islets and a reduced granulation and number of B cells. Quite coincidentally, Platt (69) described exocrine pancreatic lesions in guinea pigs infected Infectious and Immune Mechanisms 77 with foot-and-mouth disease. Certainly the above is sufficient to deem additional study as badly needed, in terms of both human and animal models. Any such studies in animals must include the spectrum of bio- chemical, physiologic, and anatomic methods available for the study of disease processes. In most cases our knowledge of pathogenesis is fragmentary. Detailed anatomic studies (histolog- ical and ultrastructural) are the only way to carefully document the pathogenesis of the disease. Hopefully some animal model will be found in which the pathogenesis resembles that of human diabetes mellitus. IMMUNE MECHANISMS IN THE ETIOLOGY AND/OR PATHOGENESIS OF HUMAN AND EXPERIMENTAL DIABETES MELLITUS Inflammatory changes in the islets in acute juvenile diabetics has been documented by Gepts (22, 23) and LeCompte (42). LeCompte and Legg (43) have documented recently two cases of "insulitis" in maturity onset diabetes. Reports of inflammatory change dating to the early studies of Opie (65) and Heiberg (30) are even more convincing in the work of Gepts (22) and Steiner (73), and are reviewed by Warren, LeCompte, and Legg (77) and Gepts (23). To summarize Gepts' opinion (22, 23) the characteristic pathologic changes in the islets of 22 cases of recent onset (acute) juvenile diabetes mellitus consist of degranulation of B cells (100 per- cent) ; hydropic change (the cytoplasmic inclusion described on p. 79 (53 percent); fibrosis (63 percent); mild insulitis (68 percent), and cytoplasmic basophilic bodies (Kornchen) are con- stantly encountered. Insulitis is defined as the presence of inflammatory cells, mainly lymphocytes in the islet proper. Even though the inflammatory cell infiltrate is mild in most instances, the presence of a mild inflammatory reaction has led numerous individuals to specu- late on allergic, immune, or autoimmune mechanisms in the etiology and/or pathogenesis of diabetes. Two converging lines of evidence have provided support for the concept of a possible immune mechanism in diabetes. First, the development of a form of experimental diabetes in animals using techniques of immunology; and secondly, the discovery of tissue specific antibodies in acute human clinical diabetes. Animal models of so-called "immunodiabetes' were explored as a consequence of finding anti- bodies to insulin in sera of diabetic patients. Dating from the work of Moloney and Coval (54), numerous studies have been done on the influence of antibodies on pancreatic islet function. Diabetic syndromes and/or islet pathology (insulitis) can develop as a direct response to the repeated injection of heterologous insulir (25,37,44,75), as well as by injecting one species with insulin antibodies made in a different species (2,3,39,49,50,79). The pathology in both situations, i.e., direct immunization and transfer of antibodies, is comparable. The animals (mice, rats, rabbits, guinea pigs, cows) evidence B cell granulation, B cell destruction, and lymphocytic infiltration in the islets. In most instances, hyperglycemia persists (i.e., chronic diabetes develops). As pointed out by numerous workers (9,39) the pathologic changes do resemble those seen in the islets of babies born to diabetic mothers. However, severe insulitis and B cell necrosis are not part of the pathology of acute human juvenile diabetes (23). The severity of the insulitis is variable in immunodiabetes. Korcdkovd, Titlbach, and 78 Diabetes Mellitus Lomsky (36) in their study of immunodiabetes in guinea pigs have described minimal, if not negligible, inflammatory changes in the islets. They also have demonstrated precipitating (circulating) antibodies in these guinea pigs injected with beef insulin. The absence of a cellular infiltrate in guinea pig islets is in direct contrast to all other reports of the islet pathology in immunodiabetes. The ultrastructural characteristics of this type of immunodiabetes has been described by Titlbach and Korcakova (74). Eleven animals were studied, and the most marked change was degranulation of B cells with masses of granular ER present in the cytoplasm. Small amounts of glycogen were seen in one animal. This particular case had B cells which by electron microscopy resembled those seen in the infectious model of guinea pig diabetes (58). In a subsequent study, Korcakova, Titlback, and Nouza (37) reported cyclophosphamide administration after and/or before onset of immunization inhibited the formation of insulin antibodies and a reduced severity of islet pathology. This study was used to confirm the concept that immune mechanisms are operable in the pathogenesis of diabetes since this cytotoxin had been used to reduce the severity of experimental thyroiditis and allergic encephalomyelitis. This type of study needs to be expanded to cover other animal models, and other antigens possibly capable of producing immunodiabetes need to be explored in addition to beef insulin. With a variable degree of insulitis now demonstrable in immunodiabetes, research in this area could be extremely profitable. Considering our present state of knowledge of immunology, and a potential future role for transplants as a possible cure for diabetes (47), our need for knowledge concerning the immunology of the endocrine pancreas should be a priority area for research. What is needed is a thorough study of immune mechanisms of all endocrines, especially the pancreatic islets. What components of normal islets are potentially immunogenic, as well as capable of producing lesions in pancreatic islets? What are the possible roles of antibodies in disturbing islet function? The list of potential specific research projects is endless, and present technology is available and adequate to answer the questions. The second factor implicating immune mechanisms in the etiology and/or pathogenesis of diabetes is the recent discovery of various tissue specific antibodies in sera of acute diabetics. Antibodies to insulin have been found in the absence of insulin therapy in numerous studies on acute diabetic subjects (10,68). Chetty and Watson tested 167 diabetic individuals not treated with insulin and 58 percent had a positive complement consumption test (antibodies to insulin) as opposed to 28 percent of controls. A second recent development is the recognition of the association of pernicious anemia, thyroiditis, and diabetes. Moore and Neilson (55), as well as Landing et al. (40), have ob- served an association of chronic thyroiditis and diabetes. Moore and Neilson reported 83 diabetic subjects with approximately 20 percent having complement fixing antibodies to thyroid and gastric mucosa. Ungar et al. (76) reported in 400 diabetics (200 insulin-dependent and 200 not insulin-dependent) an incident rate for pernicious anemia of 4 percent in insulin-dependent diabetics, whereas the incident rate was 0 percent in non-insulin-dependent diabetics. In this same study, 28 percent of insulin-dependent diabetics were found to have antibodies to gastric parietal cells. Irvine et al. (32) studied 1054 diabetics and 871 controls. They found an in- creased incidence of antibodies to thyroid cell cytoplasm and gastric parietal cell cytoplasm in insulin dependent diabetics. Infectious and Immune Mechanisms 79 Perhaps the most exciting recent development is that described by Nerup et al. (62) in demonstrating an "organ-specific, species-specific, antipancreatic hypersensitivity of the cellular type'" in diabetics of short duration in the absence on insulin therapy. These studies were done on pig pancreas where the pancreatic ducts had been ligated to produce atrophy of acinar tissue, and the sera of five out of six acute clinical diabetics inhibited leukocyte migration into the pancreas. A similar concept was used by Nerup (60, 61) to prove that hypersensitivity may play a role in the pathogenesis of idiopathic Addison's disease. The existence of Schmidt's syndrome (thyroid and adrenal insufficiency) with coexistent diabetes in 10 out of 15 cases (8) would certainly strengthen the argument that these endocrine deficiencies can be autoimmune in mechanism. If idiopathic Addison's disease, pernicious anemia, and chronic thyroiditis can be conceptu- alized as autoimmune diseases, and autoantibodies are present in acute clinical diabetes, then the opinion of Lancet in an editorial (15) is an understatement--more research in this area is badly needed and long overdue. Patients with diseases suspected of being autoimmune in nature should be studied in detail and repeatedly for the possible onset of asymptomatic and/or overt diabetes. These patients should also be screened (as described subsequently) for associated or antecedent infectious events. CONTAGIOUS DIABETES MELLITUS IN GUINEA PIGS For the past several years our group at Hershey has been studying a contagious model of diabetes mellitus in guinea pigs (57, 58, 59). Since this model has only been the basis of one scientific report, our findings to date will be briefly summarized. An original group of 18 Abbyssinian guinea pigs was obtained from a local fancier. Several members of her colony were found to be diabetic (hyperglycemic, glycosuric, and had demonstrable lesions in the pancreatic islets). In periods of time from 6 weeks to 3 months, many (approxi- mately 60 percent) white Hartley guinea pigs purchased from Perfection Breeders and brought into our colony began to have glycosuria, hyperglycemia, and all diabetic animals autopsied in the acute phase of the disease had characteristic lesions in the pancreatic islets. In addition to elevated GTT's, the serum triglycerides were elevated, as was aorta cholesterol, even though serum cholesterol was normal (41). The acute pancreatic lesion consists of: 1) degranulation of B cells as evidenced by a reduced aldehyde fuchsin stainability in paraffin sections, a reduced number of B granules, and also a reduced electron opacity of individual granule cores as seen in the electron microscope; 2) an increase in cytoplasmic masses of granular ER also seen at the light microscopic level as basophilic bodies (Kornchen) ; 3) a cytoplasmic inclusion consisting of masses of material re- sembling glycogen and admixed with B granule cores recognized in paraffin sections as "hydropic degeneration;" 4) a striking fibrosis in the vascular stroma of the islets; and 5) the presence of elongated fibroblast processes between the capillaries and the endocrine cells of the islets. The kidneys, muscle capillaries, and other organs of acutely ill animals are entirely normal. At no time have any infectious agents been observed in the islets, and inflammation or insulitis was never present. The cytoplasmic inclusion in paraffin sections frequently dropped out of the section giving rise to the appearance of '"hydropic degeneration." The inclusion is periodic acid Schiff (PAS) positive and removed by diastase indicating the presence of carbohydrate moieties that stain like 80 Diabetes Mellitus glycogen. These pathologic changes closely resemble those enumerated by Gepts (22) in acute juvenile diabetes in man (cited on p. 73), Asymptomatic as well as overt diabetics are seen in this animal model (58). All had elevated GTT's at 1 and/or 4 years, At autopsy the pancreas was nearly normal; however, renal glomerular lesions were striking. The glomeruli in PAS-stained sections, as well as in electron micrographs, evidenced marked thickening of the basal lamina, especially prominent around the mesangial cells. Small nodules of PAS-positive material were present, associated with peripheral capillary loops. Although these nodules were not as massive as in those encountered in human Kimmelstiel-Wilson disease of the kidney, they were distinctive enough to deserve further study. The few (six) chronic animals (diabetes of 1 to 3 years' duration) show a consistent pan- creatic lesion and a variable renal and vascular lesion. The pancreatic lesion consists of severe hyalinization and fibrosis of the islets, a variable degre of B cell degranulation, and variable prominence of the cytoplasmic inclusion. The renal lesions are so variable as to deserve further study prior to any scientific report. Capillary basal laminae in striated muscle are quantitatively thickened (70). This animal model is clearly contagious in nature, but an infectious agent has never been seen in the pancreatic islets. It could well involve immune mechanisms in the pathogenesis of both pancreatic as well as renal/vascular lesions. SUMMARY We have presented evidence that diabetes mellitus in man may be associated with infectious events prior to the onset of clinical diabetes. Certain animal models of viral infections can cause necrosis and inflammation of islets. The guinea pig as a contagious model perhaps also involing immune mechanisms still awaits elucidation as to the nature of the transmissible agent. Certainly, the immune state of some diabetics suggests an association that could be part of the pathogenesis of human disease. Immune factors involving renal disease have been well-documented in the case of glomerulonephritis. Since diabetics presently do not usually die from their pancreatic disease, the importance of the pathogenesis of the other organ systems assumes an added importance. One potential spin-off from the current surge of interest in viral oncology will be methods and concepts of studying infectious agents that may not fit our classical con- cepts of a ''viral infection,'" i.e., poliomyelitis destruction of anterior horn cells. With the current interest in transplantation and cancer, the field of basic immunology has exploded. The present arguments would suggest several immediate courses of action. 1) Basic research on areas touched in this review needs to be strengthened including more research on the following: a. A study of the normal developmental biology of islets (a potential value also for transplantation applications). b. Study of altered states of endocrine secretion in animal models using known physio- logic parameters to alter islet cell function and follow cellular alterations in the islet. We do not have an adequate knowledge of the repertoire of possible normal cellular activity in pancreatic islets. c. Basic pathologic (light and electron microscopic) studies of human endocrine pancreas, in both diabetics and nondiabetics. Infectious and Immune Mechanisms 81 d. Application of techniques of modern virology and immunology to study of diseases of the endocrine organs, including diabetes. e. Research efforts involving animal models in which infectious and/or immune mecha- nisms have been implicated need to be expanded. 2) A new approach to handling of human diabetes: a. New cases of overt diabetes should be studied exhaustively, using the resources of academic health centers on a regional basis. Detailed histories of prior infectious events, antibody titers to those agents thought to be associated with the onset of diabetes (mumps, rubella, Coxsackie viruses), the presence of organ specific anti- bodies, relationships to other clinical entities thought to be immune in nature, all should be studied in detail. b. Clinical diabetics should be followed with periodic exhaustive clinical, laboratory, and pathologic study on a long-term basis, again with the resources of the academic health center. The pathologic changes in kidney need to be studied temporally by light and electron microscopy. Those cases of justifiable biopsy of the pancreas (cysts, trauma, and cancer) need to be studied by light and electron microscopy, collecting and sharing this relatively rare tissue regionally. Such studies would provide better documentation of the pathogenesis of human diabetes, both in the endocrine pancreas and neurovascular systems. c. Families of known diabetics should be examined for potential diabetes. These individuals should be exhaustively studied and followed, watching for infections and changes in their immune system. I would predict a possible explosion in knowledge regarding the etiology, pathogenesis, and management of human diabetes mellitus in the next decade. Certainly the concepts of infection and immune mechanisms provide a new basis for approaching the study of diabetes. Our research efforts of the past 40 years have been, with justification, weighted in the direction of learning about the effects of insulin in physiologic and biochemical terms. With a technology available to answer new types of questions regarding diabetes, the time is ripe for an expanded research effort on the part of the biomedical community. REFERENCES 1. Adams, SF 1926. The seasonal variation in the onset of acute diabetes. Arch Int Med 37:861-864. 2. Armin, J, RT Grant, and PH Wright 1960a. Acute insulin deficiency provoked by single injections of anti-insulin serum. J Physiol 153:131-145. 3. Armin, J, RT Grant, and PH Wright 1960b. Experimental diabetes in rats produced by parenteral administration of anti-insulin serum. J Physiol 153:146-165. 4. Barboni, E, and I Mannochio 1962. Alterazionia pancreatiche in bovini con diabete mellito post-aftoso. Arch Vet Ital 13:477-489. 5. Barboni, E, I Mannochio, and G Asdrubali 1966. Observations on diabetes mellitus associated with experimental foot-and-mouth disease in cattle. Vet Ital 17:362-368. 82 Diabetes Mellitus 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. Brown, EE 1956. Infectious origin of juvenile diabetes. Arch Pediat 73:191-198. Burch, GE, C Tsui, JM Harb, and HL Colcolough 1971. Pathologic findings in the pancreas of mice infected with Coxsackie virus B4. Arch Int Med 128:40-47. Carpenter, CC, N Solomon, SG Silverberg, T Bledsoe, RC Northcutt, JR Klinenberg, IL Bennett, AM Harvey 1964. Schmidt's syndrome (thyroid and adrenal insufficiency): A review of the literature and a report of 15 new cases including 10 instances of coexistent diabetes mellitus. Med (Balt) 43:153-180. Cerasi, E, and R Luft 1972. Clinical diabetes and theories of pathogenesis. In Handbook of Physiology, Sect. VII, Vol. I. DF Steiner and N Freinkel (Eds.), Washington, Am Physiol Soc, pp 627-640. \ Chetty, MP, and KC Watson 1965. Antibody-like activity in diabetic and normal serum measured by compliment consumption. Lancet 1:67-69. Craighead, JE 1972. Workshop on viral infection and diabetes mellitus in man. J Inf Dis 125:568-570. Craighead, JE, and MF McLane 1968. Diabetes mellitus: Induction in mice by encephalomyo- carditis virus. Science 162:913-914. Craighead, JE, and J Steinke 1971. Diabetes mellitus-like syndrome in mice infected with encephalomyocarditis virus. Am J Path 63:119-134. Dalldorf, G, and R Gifford 1952. Adaptation of group B Coxsackie virus to adult mouse pancreas. J Exp Med 96:491-497. Editorial: Autoantibodies in diabetes. 1970. Lancet 2:193-194. Editorial: Coxsackie viruses and diabetes. 1971. Lancet 2:804. Fechner, RE, MG Smith, and JN Middlekamp 1963. Coxsackie B virus infection in the newborn. Am J Pathol 42:493-503. Forrest, JM, MA Menser, and JA Burgess 1971. High frequency of diabetes mellitus in young adults with congenital rubella. Lancet 2:332-334. Forrest, JM, MA Menser, and JD Harley 1969. Diabetes mellitus and congenital rubella. Pediatrics 44:445-446. Gamble, DR, ML Kinsley, MG Fitzgerald, R Bolton, and KW Taylor 1969. Viral antibodies in diabetes mellitus. Brit Med J 3:627-630. Gamble, DR, and KW Taylor 1969. Seasonal incidence of diabetes mellitus. Brit Med J 3:631-633. Gepts, W 1965. Pathology and anatomy of the pancreas in juvenile diabetes mellitus. Diab 14:619-633. Gepts, W 1972. Pathology of islet tissue in human diabetes. In Handbook of Physiology, Sect. VII, Vol. I. DF Steiner and N Freinkel (Eds.), Washington, Am Physiol Soc pp 289-303. Grishaw, WH, HF West, and B Smith 1939. Juvenile diabetes mellitus. Arch Int Med 64:787- 799. Grodsky, GM, R Feldman, WE Toreson, and JC Lee 1966. Diabetes mellitus in rabbits immunized with insulin. Diab 15:579-585. Gunderson, E 1927. 1s diabetes of infectious etiology? J Infect Dis 41:197-202. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. Infectious and Immune Mechanisms 83 Hadden, DR, JH Connolly, DA Montgomery, and JA Weaver 1972. Coxsackie B virus and diabetes. Brit Med J 4:729. Harris, HF 1899. A case of diabetes mellitus quickly following mumps. Boston Med Surg J 140:465-469. Harris, HJ 1955. Brucellosis and diabetes mellitus. N Y State J of Med 55:1183-1185. Heiberg, KA 1911. Studien Uber die pathologischen-anatomischen grundlage des diabetes mellitus. Virch Arch Path Anat Physiol U Klin Med 204:175-189. Hinden, E 1962. Mumps followed by diabetes. Lancet 1:1381. Irvine, WJ, BF Clarke, L Scarth, DR Cullen, LJP Duncan 1970. Thyroid and gastric autoimmunity in patients with diabetes mellitus. Lancet 2:163-168. John, HJ 1934. The diabetic child: Etiologic factors. Ann Int Med 8:198-213. Johnson, GM, and RB Tudor 1970. Diabetes mellitus and congenital rubella infection. Am J Dis Child 120:453-455. King, RC 1962. Mumps followed by diabetes. Lancet 2:1055. Korcakova, L, M Titlbach, and R Lomsky 1972. Immuno-diabetes in the guinea pig I. Light microscopy. Z mikr -anat Forsch 85:85-97. Korcakova, L, M Titlbach, and K Nouza 1972. Cyclophosphamide inhibition of insulin anti- body production, insulin resistance and experimental immunodiabetes. Acta Diab Lat 1X:924-957. Kremer, HV 1947. Juvenile diabetes as a sequel to mumps. Am J Med 3:257-258. Lacy, PE, and PH Wright 1965. Allergic interstitial pancreatitis in rats injected with guinea pig anti-insulin serum. Diab 14:634-642. Landing, BH, MD Petit, RL Wiens, H Knowles, and GM Guest 1963. Antithyroid antibody and chronic thyroiditis in diabetes. J C1 Endocr Metab 23:119-120. Lang, CM, and BL Munger 1973. Diabetes mellitus in the guinea pig (in preparation). LeCompte, PM 1958. '"Insulitis'" in early juvenile diabetes. Arch Pathol 66:450-457. LeCompte, PM, and MA Legg 1972. Insulitis (lymphocyte infiltration of pancreatic islets) in late onset diabetes. Diab 21:762-768. LeCompte, PM, J Steinke, JS Soeldner, and AE Renold 1966. Changes in the islets of Langerhans in cows injected with heterologous and homologous insulin. Diab 15:586-596. Lee, JC, GM Grodsky, J Caplan, and L Craw 1969. Experimental immune diabetes in the rabbit. Am J Pathol 57:597-616. Leon, AP, and N Aguirre 1953. Brucellosis, diabetes, and liver deficiency. Am J Publ Health 43:539-541. Leonard, RJ, A Lazarow, and OD Hegre 1973. Pancreatic islet transplantation in the rat. Diabetes 22:413-428. Levy, NL, and AL Notkins 1971. Viral infections and diseases of the endocrine system. J Inf Dis 124:94-103. Logothetopoulos, J, and EG Bell 1966. Histological and autoradiographic studies of the islets of mice injected with insulin antibody. Diab 15:205-211. Logothetopolous, J 1968. Electron microscopy of the pancreatic islets stimulated by insulin antibody. Canad J Physiol Pharm 46:407-410. 84 Diabetes Mellitus 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. McCrae, WM 1963. Diabetes mellitus following mumps. Lancet 1:1300-1301. Melin, K, and B Ursing 1958. Diabetes mellitus som komplikatin till parotitis epidemica. Nordisk Med 60:1715-1717. Messaritakis, J, C Karabula, C Kattamis, N Matsaniotis 1971. Diabetes following mumps in siblings. Arch Dis Child 46:561-562. Moloney, PJ, and M Coval 1955. Antigenicity of insulin. Diabetes induced by specific antibodies. Biochem J 59:179-185. Moore, JM, and JME Neilson 1963. Antibodies to gastric mucosa and thyroid in diabetes mellitus. Lancet 2:645-647. Munger, BL 1972. The biology of secretory tumors of the pancreatic islets. In Handbook of Physiology, Sect. VII, Vol. I. DF Steiner and N Freinkel (Eds.), Washington, Am Physiol Soc, pp 305-314. Munger, BL, and CM Lang 1972. Diabetes mellitus in guinea pigs: An infectious model with pathologic changes resembling its human counterpart. Diabetes 21 (Suppl 1):338. (Abstract) Munger, BL, and CM Lang 1973. Renal and pancreatic lesions in diabetic guinea pigs in the absence of glycosuria. Diabetes 22 (Suppl .1):307. (Abstract) Munger, BL, and CM Lang 1973. Spontaneous diabetes mellitus in guinea pigs--The acute cytopathology of the islets of Langerhans. Lab Invest (in press). Nerup, J, and G Bendixen 1969. Anti-adrenal cellular hypersensitivity in Addison's disease II. Correlation with clinical and serological findings. Clin Exp Immunol 5:341-353. Nerup, J, V Anderson, and G Bendixen 1970. Anti-adrenal cellular hypersensitivity in Addison's disease IV. In vivo and in vitro investigations on the mitochondrial fraction. Clin Exp Immunol 6:733-739. Nerup, J, 00 Andersen, G Bendixen, J Egeberg, JE Poulsen 1971. Antipancreatic cellular hypersensitivity in diabetes mellitus. Diab 20:424-427. Oberdisse, K (ed.) 1967. Brook Lodge workshops on spontaneous diabetes in laboratory animals. Diabetologia 3:63-286 Oberdisse, K (ed.) 1970. Brook Lodge workshops on spontaneous diabetes in laboratory animals. Diabetologia 6:153-372. Opie, EL 1900-01. The relation of diabetes mellitus to lesions of the pancreas. J Exp Med 5:527-540. Pappenheimer, AM, LJ Kunz, and S Richardson 1951. Passage of Coxsackie virus (Connecticut- 5 strain) in adult mice with production of pancreatic disease. J Exp Med 94:45-64. Patrick, A 1924. Acute.diabetes following mumps. Brit Med J 2:804. Pav, Jy Z Jezkova, and F Skrha 1963. Insulin antibodies. Lancet 2:221-222. Platt, H 1958. Observations on the pathology of experimental foot-and-mouth disease in the adult guinea pig. J Path Bact 76:119-131. Radin, AM, and BL Munger. Muscle capillary basement membrane changes in spontaneous diabetes mellitus in guinea pigs. In preparation. Rich, A 1951. The Pathogenesis of Tuberculosis, 2nd Ed. Springfield, CC Thomas. Robertson, JS 1954. The pancreatic lesion in adult mice infected with a strain of pleur- odynia virus I. Electron microscopical investigation. Austr J Exp Biol Med 32:393-410. 73. 74. 75% 76. 77. 78. 79. is Infectious and Immune Mechanisms 85 Steiner, H 1968. Insulitis beim perakuten diabetes des kindes. Klin Woseh 46:417-421. Titlbach, M, and L Korcdkovd 1972. Immunodiabetes in the guinea pig. II. Electron microscopy. Z mikir-anat Forsch 85:199-217. Toreson, WE, JC Lee, and GM Grodsky 1968. The histopathology of immune diabetes in the rabbit. Am J Pathol 52:1099-1115. Ungar, B, AE Stocks, FI R-Martin, S Whittingham, and IR Mackay 1968. Intrinsic-factor antibody, parietal cell antibody, and latent pernicious anemia in diabetes mellitus. Lancet 2:415-417. Warren, S, PM LeCompte, and MA Legg. The Pathology of Diabetes Mellitus. Philadelphia, Lea and Febiger, 1966. White, P 1927. The potential diabetic child. JAMA 88:170-171. Wright, PH 1961. The product of experimental diabetes by means of insulin antibodies. Am J Med 31:892-900. In view of the very rapidly evolving developments in this field, the reader's attention called to the following additional recent reviews: Maugh, TH 1975. Research News, Diabetes: Epidemiology suggests a viral connection. Science 188:347-351. Maugh, TH 1975. Research News, Diabetes (II): Model systems indicate viruses a cause. Science 188:436-438. Editorial: Autoimmune diabetes mellitus. 1974. Lancet 2:1549-1551. Steinke, J, and KW Taylor 1974. Viruses and the etiology of diabetes. Diabetes 23:631-633. FACTORS INFLUENCING DEVELOPMENT OF THE DIABETIC STATE Chapter 8 Chapter 9 Environmental Factors Influencing the Development of the Diabetic State Ronald A. Arky The Role of the Neuroendocrine System in the Development of Diabetes Mellitus Daniel Porte, Jr. 89 106 any a vs =. nla I - = i PIR TAR ie oh “i ) Wn hana nm 13 Le - xn = oo B Fle A - fy oa $ dy, od B em sh Ff oA "ENVIRONMENTAL FACTORS INFLUENCING THE DEVELOPMENT OF THE DIABETIC STATE Ronald A. Arky Controversy persists regarding the relative importance of hereditary and environmental factors in the development of the diabetic state. Consensus acknowledges that both factors are relevant, that in certain populations one factor may predominate, while in other populations the other factor predominates. This brief review will address itself to specific environmental factors that are known to influence the appearance of the diabetic state. Among these are: 1) nutritional factors, 2)infec- tion, 3) pregnancy, and 4) pharmacological agents. NUTRITIONAL FACTORS Since the original description of diabetes mellitus, nutritional principles have been considered as factors in both the pathogenesis and treatment of the disease. Not until 1788 did Rollo (10) describe a sound foundation for dietary treatment of diabetes. Over the last century, it has become apparent that a paradox exists regarding diabetes mellitus and the status of an individual's nutri- tion. States of excessive caloric intake (obesity) and states of caloric deprivation (starvation) are frequently associated with abnormal carbohydrate tolerance. a) Starvation i. Acute starvation. Claude Bernard (9) described ''starvation diabetes' when he noted: "It is possible to make an animal diabetic if carbohydrate is given under certain circumstances namely if sugar is given to it after 24 to 36 hours of fast." Lehmann (64) corroborated these findings when he observed gluco- suria in dogs that renewed feeding after prolonged fast. Bange (4) made similar observations in starved human subjects and coined the term ''pseudodiabetes.'" DuVigneud and Karr (28) demonstrated that the severity and duration of the carbohydrate intolerance that followed a period of starvation correlated with the length of starvation. Ingle (49) induced starvation diabetes in rodents fasted for 10 days and observed that insulin administered to such animals only slightly improved the defect. He concluded that "starvation diabetes' is not caused by hypoinsulinism. This conclusion was corroborated by Unger et al. (105) using a radioimmunoassay to measure insulin and to demonstrate that although hyperinsulinism accompanies the glucose intolerance in healthy subjects starved for three days, it was associated with an initial delay in insulin release. Yalow and her associates (112) observed that in nonobese subjects fasted for two and one-half days, the glucose intolerance (after an oral glucose load) occurred in the setting of "diabetic-type hyperinsulinism;" however when subjects of similar body build were fasted for five and one-half days the "peak insulin concentration never significantly exceeded" the peak value observed in the pre-fast study. Cahill and his colleagues (17) fasted six healthy subjects for eight days and observed that an intravenous load of glucose was removed from the circulation at a greatly diminished rate when con- trasted to the rate observed before fasting. These studies showed that the insulin response to glucose after fasting was prompt but of lesser magnitude than that prior to the deprivation of calories. These workers also demonstrated that the effectiveness of endogenous insulin was markedly decreased after starvation. 89 90 Diabetes Mellitus Obese, nondiabetics respond to acute total starvation in a manner similar to normal weight subjects. Several studies (Table 1) indicate that in overweight subjects with normal carbohy- drate tolerance prior to fasting, starvation impairs the ability to handle an oral or intravenous challenge with glucose. Such impairment generally occurs without discernible alterations in the magnitude of the insulin response, although frequently the peak insulin levels observed after starvation occur later in time than in the pre-fast test. (Figure 1). TABLE 1. Effects of Acute Starvation on Glucose Tolerance and Insulin Response in Obese Nondiabetics Number of Duration Glucose Tolerance Author Patients of Fast Type/Result vs. Prefast Insulin Response Beck et al. 4 4-14 days Oral - 100 Gm Delayed peak. Mean insulin Impaired response unchanged from pre- fast value. Genuth, S. 6 Minimum I.V. - 25 Gm Peak response slightly (1966) 6 days Disappearance rate fell. delayed. Integrated area K from 1.30 to .84 %/min. unchanged. Sussman, K. 12 5-14 days Oral - 100 Gm Delayed peak. (1966) Impaired Jackson, 'L.M.D, 5 2 1/2-18 Oral - 50 Gm Peak generally delayed. et ‘al, (1968) weeks Deteriorate 0 and 30 minute levels lower than prefast. Tzagournis, M. 7 14 days Oral - 100 Gm Delayed peak. Diminished. et al. (1970) Deteriorate 1/G ratio at one hour. Jackson, R. A. 7 14 days Oral - 100 Gm Delayed peak. et al. (1972) Levels at 90 and 180 minutes higher after starvation 30 —o— Before Fastir 140 1 —o— Before Fasting e —o— After Fasting --0-= After Fasting 00 b mb appr BE X ks wE mF ggs 100 ggs “[ i 237 wf rss 2 gm. 7 I. ® 120 | 7 Sw x [1] Vd T ~ 0 Lo 150 i of 2s 22% sf ahd. = 3 5 lo >= 4 . SER “23 5k 4 50 = ok * 0 a 12r 12 < - o EE Sus @ = 208 S<8% 0s 328 Le Boma, ES 28 : BS ul TET 0.4 x Ee 12 By. * Ew géz 4 Bie, 5. op A 58% & Ee 27h 382 oral . & 0 8 0 4 wv wv Rebellion ibid hh Lilies) 0 120 20 %0 0 120 20 360 MINUTES VINES FIGURE 1. and diabetics. the left. tested with 100 gram loads (52). Diabetes The response (mean + SE) to oral glucose loading in obese non- diabetics before and after starva- tion. Effects of starvation on glucose tolerance in obese normals Results of the normal (nondiabetic) subjects are on Fasts lasted for 14 days and oral glucose tolerance was Reprinted with permission from the American Diabetes Assn. The response (mean = SE) to oral glucose loading in obese diabetics before and after starvation. 20:214-227,The Journal o Environmental Factors 91 Prolonged starvation produces variable changes in the glucose tolerance of obese, mild diabetics. Reports of improvement (Figure 1) (5,51,90), of improvement and deterioration (112), and no change (38) have been made. These studies point up the heterogeneity of obese maturity onset diabetics as a group. Sweeney (101) emphasized that it is the deficiency of dietary carbohydrate that leads to 'starva- tion diabetes." Feeding medical students a ''pure fat! diet causes an alteration in glucose tolerance analogous to that induced by total starvation. A number of studies (Table 2) have attempted to define the minimal dietary carbohydrate required to inhibit deterioration in glucose tolerance. While Wilkerson et al. (110) suggest that a 150 gram carbohydrate intake is as effective as Conn's (23) 300 gram carbohydrate diet in preventing alterations in glucose tolerance, some questions remain as to the minimum dietary carbohydrate needed tc prevent alterations in glucose tolerance. TABLE 2. Studies Comparing the Effect of Dietary Carbohydrate on Glucose Tolerance Period of Carbohydrate Content Investigator Preparation (Grams) of Trial Diets Comments ‘Himsworth 5 days 55 vs 600+ Use 50 Gm oral load. Low carbohy- (1933) drate diet impairs tolerance. High diet impractical. Conn 3 days-300 Gm 20 vs 300 Used 1.75 Gm glucose per Kg, Felt (1940) 5 days- 20 Gm 300 Gm carbohydrate needed in prepar- atory stage. Showed misdiagnosis on low carbohydrate intake. Irving and 4 days 100 vs 300 Used 100 Gm glucose. Lower carbo- Wang (1954) hydrate caused earlier high levels. Two hour levels were equivalent. Wilkerson 4-5 days Usual vs 20 vs 150 Deterioration on 20 Gm intake returns et al. (1960) to "usual" after four days on 150 Gm. Hales and Randle 5 days Usual (>200) vs <50 Performed both 50 and 100 Gm oral (1963) tolerance tests. Deterioration on <50 Gm diet. ii. Chronic starvation. That acute starvation induces an abnormal tolerance to glucose in man and other species is well established; however the effects of prolonged chronic starvation and malnutrition are not completely defined. Kwashiorkor is a chronic form of starvation that develops in youngsters ingesting diets poor in protein yet containing calories from nonprotein sources (30). This form of starvation is accompanied by glucose intolerance and a diminished insulin response after a glu- cose challenge (3,13,96). Marasmus develops in children whose diets are deficient in both protein and total calorie content. These children have relatively normal low fasting glucoses and lower normal plasma free fatty acids. However, carbohydrate tolerance has been reported as normal (13, 40). Chronic starvation is noted in adult populations in the form of anorexia nervosa, a condition reported to be accompanied by normal glucose tolerance (26,87). b) Overfeeding - obesity. i, Induced obesity. In contrast to starvation hyperphagia or forced overfeeding also induces alterations in carbohydrate tolerance. Hofmeister (46) observed that overfeeding dogs with carbohydrate caused glucosuria. Lesions in the ventromedial nucleus of the rat's hypothalmus caused hyperphagia and 92 Diabetes Mellitus consequently an impaired carbohydrate tolerance (14). Similarly hyperphagic monkeys with hypothalamic lesions demonstrate a high incidence of postprandial hyperglycemia and glucosuria (42). Keys et al. (61), Mann et al. (69), and most recently Sims et al. (94) have overfed subjects and assessed metabolic changes. The last group of investigators overfed two groups of subjects, four University of Vermont students and then later a group of prisoner volunteers at the Vermont State Prison. Each subject ingested two to three times his normal intake of calories; the prisoners reached an average of 26 percent above their initial lean weight and their adipose tissue mass increased by 70-100 percent. Significant reduction in the oral and intravenous glucose toler- ance followed weight gain although the oral tolerance curve and disappearncce curve after intravenous glucose remained within the "accepted range'" (93). After the gain in weight the pancreatic secretion of insulin was greater than that observed prior to overeating, even though the glucose challenge was comparable in both instances. Mahler (67) performed a similar experiment with healthy male students who ingested an additional 50 percent of their normal caloric intake either at one sitting (guzzling) or at hourly intervals throughout the day. All subjects gained weight: the 'guzzlers' gained more than the 'mibblers." Fasting blood glucose increased significantly at the end of the second week of 'guzzling'" but did not change in the 'mibblers." However, glucose tolerance (50 grams) as measured by the area above the baseline value did not change significantly during either type of feeding. During overeating mean fasting insulin levels increased from 11.2 + 2.8 microunits/ml to 32.5 +5 microunits/ml after one week of '"'guzzling" but were unchanged after 'mibbling." Further studies (48) showed that although basal insulin increased nearly 50 percent in five volunteers after overfeeding had induced a 15-25 percent gain in weight, the fasting arterial glucose levels were unchanged. However, those circulating amino acids that are usually governed by insulin, e.g., leucine, isoleucine, tyrosine, phenylalanine, and valine increased by 18 to 25 percent. This "peripheral resistance to insulin after overfeeding was further substantiated by studies in which exogenous insulin injected interarterially after weight gain had a diminished effectiveness when contrasted to the effectiveness prior to the weight gain. ii. Spontaneous obesity. l These observations on "overfeeding'" or induced obesity emphasize several relationships between overweightness and disordered carbohydrate metabolism. Three salient facts that underlie current concepts are: a. The incidence of diabetes mellitus is high among the obese population. Among 4596 diabetics 78.5 percent of the males and 83.3 percent of the females were over- weight at the time of their maximum weight (70). A study of 100 obese men and women who were at least 25 percent or heavier than ideal body weight showed that 22 percent had elevated fasting blood glucoses and 58 percent had impaired glucose tolerance tests (99). Yet not all longstanding obese subjects develop diabetes and the interplay of obesity with the multiplicity of other pathogenic factors requires clarification. be Obesity is accompanied by insensitivity to insulin. Insensitivity of muscle and adipose tissue to insulin (82), hypertrophy of the pancreatic islets (73), higher basal insulin levels (2), and exaggerated insulin to glucose and other insulin secretogogues (16,60,77) all characterize obesity and theoretically can be linked into a cycle that eventuates into diabetes mellitus. In obese mice high insulin Environmental Factors 93 response progresses to an insulin deficient type of diabetes (78). The "messenger' that transmits to the beta cell that "obesity is present and more insulin is needed" defies definition at this time. Whether the elevated plasma levels of branch chain amino acids (valine, leucine, isoleucine, etc.) in obese, hyperinsulinemic subjects (34) provide this signal remains speculative. Moreover, whether the basic insulin resistance of tissues of the obese can be attributed solely to the insulin insensitivity of oversized adipose cells (88) is debatable. Cc. Weight loss may reverse some of the metabolic derangements of obesity. Newburgh and Conn (72) showed that weight loss corrected the abnormal glucose tolerance in 90 percent of a group of middle-aged obese patients. Weight loss induces improvement in glucose and tolbutamide tolerance in spite of lower basal insulin levels and reduced insulin response to various stimuli (32,59,95). c) Nature of dietary carbohydrate. Does the nature of dietary carbohydrate play a role in the pathogenesis of diabetes mellitus? Arguments pro and con can be mustered, so that it is apparent that a dogmatic answer is unavailable. Since the recognition of diabetes mellitus as a "disorder of sugar metabolism," there has always been a popular notion that ''too much sugar might induce the problem.' While this proposition has been in part reputed, a number of epidemiological studies reignite general interest in the problem. Each of these studies considers the generic term ''carbohydrate' to consist of two general classes: a) the "natural, unconcentrated carbohydrate as exists in grain, potatoes and fruit and b) the "unnatural, concentrated carbohydrates' as exemplified by refined sugar and flour. A brief summary of these epidemiological studies and the populations involved: i. Yemenites. Cohen (20, 21) cites the prevalence of diabetes mellitus among new Yemenite immigrants to Israel as 0.06 percent, while that in Yemenite settlers who have been in Israel for 25 years as 2.9 percent. He attributes this marked increase in prevalence to the high sucrose content (25-30 percent of total carbohydrates in the Israeli diet is sucrose) of the old settlers' diet as con- trasted with the high starch, low to absent sucrose content of the diet eaten in Yemen. While "old settlers' are heavier and report a slight increase in caloric consumption compared to immigrants, Cohen feels that the substantial appearance of diabetes after acclamation to Israeli dietary habits is related to diminished consumption of bread and dietary polysaccharides which are replaced with increased consumption of disaccharides as sucrose. Cohen and his associates (22) attempt to demon- strate '"better' glucose tolerance after preparation with a high polysaccharide diet than after a disaccharide diet. ; ii. Indians. Cleave and Campbell (19) note that the overall prevalence of diabetes in India is less than one percent, yet among Indians in Natal the prevalence ranges from 2.3 percent ("barrack dwellers") to 5.2 percent (''village dwellers"). They attribute this 'veritable explosion of dia- betes..." among peoples of the same origin to the higher consumption of 'unnatural refined sugars" in Natal. Evidence of a similar trend is observed among urbanized Zulu as contrasted with tribal Zulus who have a low prevalence of diabetes. 94 Diabetes Mellitus iii. ‘Great Britain. Concerned with the increasing frequency of fatal coronary attacks and the increasing prevalence of atherosclerosis and diabetes in Britain, Yudkin (114, 115) cites the massive consump- tion of refined sugar in present-day England (120 pounds per head per year). This he contrasts to the English diet of the 1800's that abounded in fruits and 'matural carbohydrates.' He implicates the high disaccharide content of the diet as a major factor in the pathogenesis of atherosclerosis and diabetes. A number of animal studies have sought to differentiate the effects of the 'two classes" of carbohydrate on glucose tolerance. Uram et al. (106) detected only minor differences in post- glucose blood sugars in rats maintained on casein-sucrose diets as compared to animals on cereal ration. Early in the century von Noorden (106) showed marked benefits from an oatmeal diet to diabetics with profound glucosuria. As emphasized by West and Kalbfleisch (108), it is extremely difficult to evaluate precisely the sugar intake of a large population; moreover, 'refined sugar" intake is associated with affluence and with an increased total caloric intake. d) Frequency of meals. The temporal intervals between the ingestion of calories influences the metabolic perturbations induced by feeding and possibly the disposition of the nutrients. Gwinup et al. (39) demonstrated in four subjects an impaired ability to handle an oral glucose load after the individuals had been on isocaloric diets fed once daily (between 4:00 and 5:00 p.m.). These workers distributed evenly in ten feedings at two-hour intervals and after three-meal-a-day pattern and found that ''guzzling' (one meal per day) produced the worse glucose tolerance. Young et al. (113) observed similar results when college students on reducing diets were studied after periods of "nibbling" and ''guzzling". Broader implications of the effects of meal frequency upon several metabolic parameters were apparent in the study of Fabry et al. (29). A survey of 379 men revealed an inverse correla- tion between the frequency of meals and the degree of carbohydrate intolerance, as well as the incidence of obesity and hypercholesterolemia; i.e., the fewer the meals eaten per day, the greater the tendency for carbohydrate intolerance. These survey data are compatible with "stuff and starve' feeding patterns in rodents (47). INFECTION FACTORS Among the several ''stress' factors that affect carbohybrate tolerance infection must be given a prominent place. While it is universally acknowledged that diabetes mellitus is aggravated in the presence of infection, the role of infectious processes in the etiology of permanent diabetes mellitus has not been established. Inflammatory lesions exert their effects upon intermediate metabolism by either involving specific tissues (e.g., pancreas, liver) or indirectly by altering the body's hormonal milieu and thus influencing the availability and utilization of substrate. Little evidence for infection as an etiological factor in diabetes has been offered in the last twenty years. John (53) and Broun (15) felt that "heredity and infection' were the two prime factors in the etiology of diabetes. In 500 cases of juvenile diabetes (53) there was a recent history of infection in 164, and in 87 diabetes appeared within 60 days of an acute febrile illness. The seasonal occurrence of diabetes (36), geographic distribution, and coincidence of upper respira- tory tract infections in newly diagnosed diabetics have been used as arguments for an etiological role of viral and bacterial infections in diabetes mellitus. Environmental Factors 95 The mumps virus may directly affect the endocrine pancreas, although the number of reported cases of diabetes mellitus immediately following mumps is small (45, 71); less than 20 appear in the English literature since the first suggestion of this relationship (43). Members of the picornavirus group have been implicated as etiological agents in diabetes mellitus. One of these viruses, the encephalomyocarditis virus, induces histiological changes in the islets of Langerhan four days after inoculation. The islets become shrunken and beta cells degranulate. Simultaneously with these changes, glucose intolerance is apparent and insulin secretion diminishes (25). Among other viruses that affect the pancreatic beta cell are foot and mouth disease virus, Coxsakie B virus (36, 37), Reo virus, and hepatitis virus. An infectious etiology of acute pancreatitis is questionable, but whatever the cause, acute pancreatitis is commonly accompanied by hyperglycemia and glucosuria. This "transient diabetes" infrequently persists, and if the pancreatitis is severe enough to destroy the bulk of islet tissue it is usually fatal (107). In the United States and Europe chronic pancreatitis is an uncommon cause of diabetes, but in such regions as East Africa (62), it may account for 13 percent of the newly diagnosed cases. Diabetes developing secondary to known bacterial infections of the pancreas or pancreatic abscesses is rare (107). In the early 1900's a number of clinical studies noted that hyperglycemia and glucosuria were frequently found accompanying acute infections. Labbe and Boulin (63) observed glucosuria in 75 percent of nondiabetics with febrile illnesses and could not relate the severity of the illness with the magnitude of the glucosuria. They speculated: "It is possible that recurrence of this transient disturbance of glucose balance creates a true diabetes more frequently than believed." Williams and Dick (111) demonstrated decreased dextrose tolerance in nondiabetics with respiratory and renal infections. In recent years radioimmunoassay techniques have permitted the measurement of peptide hormone fluxes in both naturally occurring and induced infections. Within 24 hours of induced infection with tularemia in man, the rate of glucose disappearance as measured by intra- venous glucose tolerance is diminished. Simultaneously the rise in insulin after glucose is greater, and the fall in insulin slower than in corresponding studies performed in the pre-infection state (92). Similar observations have been made following the induction of a self-limited febrile illness, Sandfly fever, in man (84). An intricate alteration in the circulating levels of humoral factors and metabolic substrate accompanies infection (8) and effects the hyperglycemia and glucosuria that is so frequently seen. Bagdade (2) has summarized current concepts on the modes of action of these hormonal components that are increased with infection and relates these actions to the resultant "transient diabetes." Recently elevated levels of glucagon have been detected in human and experimental infections (86); the diabetogenic influence of this peptide may augment the actions of increased glucocorticoids and catecholamines that accompany the "stress' of infection. PREGNANCY Pregnancy represents a major diabetogenic stress: a) it places a progressive demand on the maternal beta cell and the insulin secretory mechanism; b) it is a state of relative insulin resistance (55). Normal women tolerate pregnancy without significant alteration in carbohydrate tolerance; women with marginal pancreatic reserve develop an abnormal glucose tolerance or gesta- tional diabetes. As many as one pregnant woman out of 116 falls into this category (74). 96 Diabetes Mellitus While the diabetes frequently clears after delivery, 28 percent of such women have frank diabetes five years later, and by 16 years 52 percent of gestational diabetics have permanent disease (O'Sullivan, personal communciation). The risk of impaired carbohydrate tolerance increases with successive pregnancies--a woman who has had five pregnancies has three times the chance of developing diabetes as a nulliparous individual (81). During the course of normal pregnancy, basal levels of insulin increase progressively and the insulin response to serious secretogogues also is greater in the pregnant than nonpregnant state (12,89,97) (Figure 2). The need for "extra" insulin in both the basal and stimulated state is proposed to stem from: a) Maternal insulin has an abbreviated biological life due to the ability of the placenta to degrade insulin (35, 80). b) Maternal tissues are resistant to the hypoglycemic effects of insulin because of circulating "antagonists" produced in part by the placenta. GLUCOSE TOLERANCE IN NORMAL PREGNANCY BLOOD GLUCOSE mg/100m! INSULIN microunits/ml 2004 200 1 Fic. 2. Blood glucose and in- sulin levels during an oral GTT in seven normal preg- nant patients in each of the three trimesters and 7-8 weeks postpartum, compared with the levels in ten gesta- tional diabetics in the third trimester. &—2 |8' Trimester ©— 20d Trimester [normols| a—a 31d Trimester ©—8 6-8 Weeks Post 0-0 314 Trimester - Diabetics T T T — v ; y 0 30 60 90 20 © 30 60 90 120 MINUTES DURING ORAL GTT. Reprinted with permission of the Journal of Endocrinology 33:521-529, 1971. Among the placental hormones that contribute to the diabetogenic effects of pregnancy are: a) : Placental lactogen (HPL). A polypeptide immunologically similar to growth hormone with contrainsulin action when given parenterally (7), HPL causes impaired glucose tolerance in subclinical and overtly diabetic patients (56). The precise site of the contrainsulin action of HPL is undetermined; the aggravation of mild diabetes caused by infusions of HPL is not accompanied by hyperinsulinism. Environmental Factors 97 b) Progesterone. When this steroid is administered to males and hysterectomized females in quantities that cause 24 hour urinary pregnanediol excretions that simulate those observed in late gestation, fasting levels of insulin and the insulin responses to glucose and tolbuta- mide are increased. Yet the oral and intravenous glucose tolerance tests are not altered (5, 57). Isolated islets of rats treated with 5.0 mgm of progesterone for 21 days were hypertrophied and had an insulin output similar to the islets from pregnant animals (24). c) Estrogens. Estradiol benzoate causes islet cell hypertrophy, increased insulin output of isolated islet cells and an increased insulin response to glucose in rodents treated for three weeks (24). The specific effects of natural and synthetic estrogens upon glucose and insulin metabolism in normal and diabetic humans are debatable (98,102), with contentions that diabetes is ameliorated or not affected by estrogens. Fluctuations in other hormones may in part account for some of the diabetogenic influences of pregnancy. Most of the elevation in serum cortisol during pregnancy is attributed to increments in the corticosteroid binding globulin and the increment in free cortisol is minimal and not felt to be an important factor in the altered maternal carbohydrate metabolism in pregnancy. Serum growth hormone levels do not increase in pregnancy and thus play no role in this insulin resistant state (58). Consensus is that the placental hormones are the chief offenders in altering substrate balance and causing the minor deviation of glucose tolerance in normal pregnancy (75). As with diabetes in general, gestational diabetes may be accompanied by anatomic as well as biochemical abnormalities. Perinatal infant mortality is higher in insulin-requiring diabetic mothers than in nondiabetics; infants of diabetic mothers are longer, heavier, and fatter than controls (76). There is substantial evidence that 'control" of the metabolic derangement in the pregnant diabetic is the most important factor in improving the fetal salvage rate. Even among gestational diabetes the fetal salvage rate was improved and weights of infants reduced when mothers were treated with small doses of insulin during pregnancy (O'Sullivan, personal communication). PHARMACOLOGIC AGENTS During the last two decades knowledge of the mechanisms involved in the synthesis, storage, and release of insulin from the pancreatic beta cell has burgeoned. Simultaneously the use of pharmacologic agents by the public has increased markedly. Agents that impede any step in the process of insulin formation or secretion will induce diabetes. An understanding of the structure and mode of action of therapeutic compounds that affect carbohydrate metabolism has become a major responsibility of practicing physicians. Historically alloxan (mesoxalyurea) is recognized as the prototype of a beta-cytotoxic agent. Although synthesized in 1818, alloxan was not discovered to possess diabetogenic action until 1943 (27), when it was shown that the agent causes necrosis of the pancreatic islets. In normal animals the sequence of alloxan's actions are: a) a marked early hyperglycemia (1-4 hours); b) hypoglycemia lasting up to 48 hours after administration; c) chronic hypoglycemia. Although alloxan has a struc- ture similar to uric acid (Figure 3), there is no evidence that abnormalities in uric acid metabo- lism cause the accumulation of alloxan. 98 Diabetes Mellitus | | | | H 0=C cc=0 OSG: "Commer ¥ a” ll c=0 HN—C ~~N 7 H CH_OH ALLOXAN URIC ACID 2 0 H Ccoow HON | /on | SON 0=C > N NN ) is N— C=0 Fre H N -NC I ALLOXANIC ACID CH, STREPTOZOTOCIN FIGURE 3. Diabetogenic agents. Uric acid is not diabetogenic, but presented to show similarity in chemical structure to alloxan. Ascorbic acid augments the diabetogenic action of alloxan, but is not itself diabetogenic. Experimentally the quinoline derivatives, 8-hydroxyquinaldine and 8-hydroxy-6 aminoquinoline are diabetogenic. These compounds have organic metal-binding properties and Kadota and Abe (54) postulate that the diabetogenic effects of these compounds stem primarily from the binding of zinc within the beta cell and subsequent interference with synthetic processes. A specific beta-cytotoxic action analogous to alloxan is observed in the broad-spectrum anti- biotic streptozotocin. This substance causes degranulation of the beta cell and disruption of the islets. The triphasic pattern in glucose levels observed after the administration of alloxan is also noted after streptozotocin. Nicotinamide, pyrazinamide and 2-deoxyglucose protect against the diabetogenic effects of streptozotocin; nicotinic acid and glutathione are ineffective in this role (85). Streptozotocin is used to produce experimental diabetes in animals and in the treatment of insulin-secreting tumors. The mode by which streptozotocin selectively destroys the beta cell is not fully understood, but it is possible that the glucose moiety of the compound (Figure 3) binds specifically to cells with glucose receptors and that the nitrosurea group causes the actual cell damage. The autonomic nervous system participates in the control of insulin secretion. Epinephrine and norepinephrine inhibit the release of insulin from glucose-stimulated beta cells (79); the diabetogenic action of these catecholamines is mediated via alpha receptors in the beta cell. The benzothiadiazines tend to suppress insulin release. Diazoxide is the most powerful of these substances (31, 91) and its hyperglucemic action is potentiated by trichlormethiazide in normal subjects and patients with insulin secreting tumors. Diazoxide, a benzothiadiazine that causes retention of sodium and no significant loss of potassium probably exerts its hyperglycemic effects directly by inhibiting release of insulin. The benzothiadiazine diuretics that induce significant negative potassium balance may exert their diabetogenic effects through the latter effect (83). Non-thiazide diuretics as chlorthalidone also induce hyperglycemia (18). Environmental Factors 99 Diphenylhydantoin in amounts that are frequently prescribed for neurological disorders inhib- its the secretion of insulin in normal subjects (68). This agent has also been utilized to treat insulin-secreting tumors. In the hamster serotonin inhibits insulin secretion (33), but stimulates release of insulin in the rabbit (103). L-asparaginase, an agent used to treat acute leukemia diminishes insulin secretion and causes glucose intolerance (109). Glucocorticoids administered in excess may induce diabetes by increasing the breakdown of protein, increasing gluconeogenesis, interfering with glucose utilization and insulin's effectiveness. FUTURE INVESTIGATION AND EVALUATION From this brief review of several environmental factors that influence the development of the diabetic state, it is evident that many questions persist. An abbreviated list of these questions is offered as the basis for future investigation and evaluation. 1. Nutritional factors. a. Obesity 1) Are there differences in physical or metabolic characteristics between the longstanding obese subjects who develop diabetes and those who do not? 2) Is a specific "humoral factor' responsible for signaling the beta cell that obesity is present? Does the signal (if present) disappear with minimal weight reduction? 3) 1s it possible to define a population that has altered a specific dietary foodstuff over the course of several years and does not exhibit weight change? All of the populations that have shifted from "natural" to "unnatural" carbohydrate have gained weight. 4) Can the relative hyperinsulinemia of obesity be altered solely by modifying the diet yet maintain isocaloric status? (i.e., substantiation of the work of Grey, N. and Kipnis, D.M., New Engl. J. Med. 211:811-916, 1969, in a large population). b. Starvation 1) At what point during starvation in normal man is the ''diabetes'" secondary to reduced insulin production as opposed to increased peripheral resistance? 2) What characteristics differentiate the maturity onset diabetics whose carbohydrate tolerance improves with starvation and those who show no improvement? 3) What factors account for variation in carbohydrate intolerance in the several "chronic starvation" states, i.e., kwashikor va. anorexia nervosa? 2. Infection a. Does a current updating substantiate the relationship between acute infections and the onset of diabetes in North American populations? b. Is carbohydrate homeostasis affected more by mild viral infections in persons with a here- ditary predisposition to diabetes vs. no hereditary predisposition? 3. Pregnancy a. Does periodic evaluation of individuals who demonstrate gestational diabetes help reduce the incidence of overt disease in later years? b. What factors determine the onset of overt diabetes in women who have had abnormal carbohy- drate tolerance during pregnancy? Undoubtedly many other questions can be raised by the data presented in the Review. The above list is presented as a starter. 100 Diabetes Mellitus REFERENCES 10. 11. 12, 13. 14. 15. 16. 17. 18. 19. 20. Bagdade, JD, EL Bierman, and D Porte, Jr 1967. The significance of basal insulin level in the evaluation of the insulin response to glucose in diabetic and non-diabetic subjects. J Clin Invest 46:1549-1557. Bagdade, JD 1971. Infections. In Diabetes Mellitus: Diagnosis and Treatment. SS Fajans and KE Sussman, Co-Editors. New York, American Diabetes Association, Inc. pp 211-215. Baig, HA, and JC Edozien 1965. Carbohydrate metabolism in kwashiorkor. Lancet 2:662-665. Bange, I 1913. Der blutzucker. Wiesbaden. Beck, P 1969. Progestin enhancement of the plasma insulin response to glucose in rhesus monkeys. Diabetes 18:146-152, Beck, P, JH Koumans, CA Winterling, MD Stein, WH Daughaday, and DM Kipnis 1964. Studies of insulin and growth hormone secretion in human obesity. J Lab Clin Med 64:654-667. Beck, P, and WH Daughaday 1967. Human placental lactogen: studies of its acute metabolic effects and disposition in normal man. J Clin Invest 46:103-110. Beisel, WR 1972, Interrelated changes in host metabolism during generalized infectious ill- ness. Am J Clin Nutr 25:1254-1260. Bernard, C 1859. Legons sur les propriétes physiologiques et les altérations pathologiques des liquides de 1'organisme. Paris. Tome 2, p 79. Best, CH 1960. Epochs in the history of diabetes. Chapter 1 in Diabetes. RH Williams, Editor. New York, Paul B. Hoeber, Inc. pp 1-13. Bjorntorp, P, K de Jounge, L Sjostrom, and L Sullivan 1970. The effect of physical training on insulin production in obesity. Metabolism 19:631-638. Bleicher, SJ, JB O'Sullivan, and N Freinkel 1964. Carbohydrate metabolism in pregnancy. V. The interrelations of glucose, insulin, and free fatty acids in late pregnancy and post partum. New Engl J Med 271:866-872. Bowie, MD 1964. Intravenous glucose tolerance in kwashiorkor and marasmus. S Afr Med J 38:328-329. Brobeck, JR, J Tepperman, and CNH Long 1943. Experimental hypothalamic hyperphagia in the albino rat. Yale J Biol Med 15:831-853. Broun, EE 1956. Infectious origin of juvenile diabetes. Arch Pediat 73:191-198. Butterfield, WJH, ME Abrams, DJB St John, and MJ Whichelow 1967. The intravenous glucose tolerance test: peripheral disposal of the glucose load in controls and diabetics. Metabolism 16:19-34, Cahill, GF, Jr, MG Herrera, AP Morgan, JS Soeldner, J Steinke, PL Levy, GA Reichard, Jr, and DM Kipnis 1966. Hormone-fuel interrelationships during fasting. J Clin Invest 45:1751-1769. Carliner, NH, JL Schelling, RP Rossell, R Okun, and M Davis 1965. Thiazide and phthalimidine- induced hyperglycemia in hypertensive patients. JAMA 191:535-537. Cleave, TL, and CD Campbell 1966. Diabetes, Coronary Thrombosis and the Saccharine Disease. Bristol, John Wright and Sons, Ltd. pp 14-56. Cohen, AM 1961. Prevalence of diabetes among different ethnic Jewish groups in Isrfel. Metabolism 10:50-58. 21. 22. 23. 24. 25. 26. 27, 28. 29, 30. 3. 32, 33. 34. 35. 36. 37. 38. 39. 40. 41. Environmental Factors 101 Cohen, AM 1963. Fats and carbohydrates as factors in atherosclerosis and diabetes in Yemenite Jews. Am Heart Journal 65:291-293. Cohen, AM, A Teitelbaum, M Balogh, and JJ Groen 1966. Effect of interchanging bread and sucrose as main source of carbohydrate in low fat diet on glucose tolerance curve of healthy slender subjects. Am J Clin Nutr 19:59-62. Conn, JW 1940. Interpretation of the glucose tolerance test. The necessity of standard preparatory diet. Am J Med Sci 199:555-564. Costrini, NV and RK Kalkhoff 1971. Relative effects of pregnancy, estradiol and progesterone on plasma insulin and pancreatic islet insulin secretion. J Clin Invest 50:992-999. Craighead, JE and J Steinke 1971. Diabetes mellitus-like syndrome in mice infected with Encephalomyocarditis virus. Am J Path 63:119-130. Crisp, AH 1965. Clinical and therapeutic aspects of anorexia nervosa - a study of 30 cases. J Psychosom Research 9:67-78. Dunn, JS and NGB McLetchie 1943. Experimental alloxan diabetes in the rat. Lancet 2:384-387. du Vigneud, V and WG Karr 1925. Carbohydrate utilization. I. Rate of disappearance of d-glucose from the blood. J Biol Chem 66:281-300. Fabry, P, J Fodor, Z Hejl, T Braun, and K Zvolankovd 1964. The frequency of meals: its relationship to overweight, hypercholesterolemia and decreased glucose tolerance. Lancet 2:614-615. FAO/WHO Joint Expert Committee on Nutrition 1953. Third Session World Health Organization Rep Ser #72. Fajans, SS, JC Floyd, Jr, RF Knopf, J Rull, EM Guntsche, and JW Conn 1966. Benzothiadiazine suppression of insulin release from normal and abnormal islet tissue in man. J Clin Invest 45:481-492. Farrant, PC, RWJ Neville, and GA Stewart 1968. Insulin release in response to oral glucose in obesity: the effect of reduction in body weight. Diabetologia 5:198-200. Feldman, J, and H Lebovitz 1970. Serotonin inhibition of in vitro insulin release from golden hamster pancreas. Endocrinology 86:66-70. Felig, P, E Marliss, and GF Cahill, Jr 1969. Plasma amino acid levels and insulin secretion in obesity. New Engl J Med 281:811-816. Freinkel, N and CJ Goodner 1960. Carbohydrate metabolism in pregnancy. I. The metabolism of insulin by human placental tissue. J Clin Invest 39:116-131. Gamble, DR, and KW Taylor 1969a. Seasonal incidence of diabetes mellitus. Brit Med J 3:631-633. Gamble, DR, ML Kinsley, MG FitzGerald, R Bolton, and KW Taylor 1969b. Viral antibodies in diabetes mellitus. Brit Med J 3:627-630. Genuth, SM 1966. Effects of prolonged fasting on insulin secretion. Diabetes 15:798-806. Gwinup, G, RC Byron, W Roush, F Kruger, and G Hamwi 1963. Effect of nibbling versus gorging on glucose tolerance. Lancet 2:165-167. Hadden, DR, and MD Beif 1967. Glucose, free fatty acid and insulin interrelations in kwashiorkor and marasmus. Lancet 2:589-592. Hales, CN, and PJ Randle 1963. Effect of low carbohydrate diet and diabetes mellitus on plasma concentrations of glucose, non-esterified fatty acid and insulin during oral glucose tolerance tests. Lancet 1:790-794. 102 Diabetes Mellitus 42. 43. 44. 45, 46. 47. 48. 49, 50. 51. 52. 53. 54. 55. 56. 37. 58. 59. 60. 61. 62. Hamilton, CL, and JR Brobeck 1965. Control of food intake in normal and obese monkeys. Ann NY Acad Sci 131:583-592. Harris, HF 1899. A case of diabetes mellitus quickly following mumps. Bost Med Surg J 140:465-469. Himsworth, HP 1933. The physiological activation of insulin. Clin Sci 1:1-38. Hindin; E 1962. Mumps followed by diabetes. Lancet 1:1381. Hofmeister, F 1890. Ueber resorption und assimilation der Ndhrstoffe. Arch Fur Exper Path u Pharmakol 26:355-370. Hollifield, G, and W Parsons 1962. Metabolic adaptations to "stuff and starve' feeding pro- gram. II. Obesity and persistence of adaptative changes in adipose tissue and liver occurring in rats limited to short daily feeding period. J Clin Invest 41:250-253. Horton, ES, P Felig, CF Runge, and EAH Sims 1972. Human experimental obesity. Basal forearm metabolism and response to intra-arterial insulin infusion. Israel J Med Sci 8:814-815. Ingle, DJ 1948. The production of experimental glycosuria in the rat. Recent Progress in Hormone Research 2:229-253. Irving, EM, and I Wang 1954. The effect of previous diet on glucose tolerance tests. Glasgow Med J 35:275-278. Jackson, IMD, MT McKiddie, and KD Buchanan 1968. The effect of prolonged fasting on carbo- hydrate metabolism: evidence for heterogeneity in obesity. J Endocr 40:259-260. Jackson, RA, M Moloney, C Lowy, AD Wright, M Hartog, TRE Pilkington, and TR Fraser 1972. Differences between the metabolic responses to fasting in obese diabetic and obese non- diabetic subjects. Diabetes 20:214-227. John, HJ 1951. Diabetes mellitus in children. A review of 500 cases. J Pediat 35:723-744, Kadota, I, and T Abe 1954. Chemical specificity of diabetogenic action of quinoline deriva- tives. J Lab Clin Med 43:375-385. Kalkhoff, R, D Schalch, JL Walker, P Beck, DM Kipnis, and WH Daughaday 1964. Diabetogenic factors associated with pregnancy. Trans As Amer Physicians 77:270-280. Kalkhoff, RK, BL Richardson, and P Beck 1969. Relative effects of pregnancy, human placental lactogen and prednisoline on carbohydrate tolerance in normal and subclinical diabetic sub- jects. Diabetes 18:153-163. Kalkhoff, RK, M Jacobson, and D Lemper 1970. Progesterone, pregnancy and augmented plasma insulin response. J Clin Endocr 31:24-28. Kalkhoff, RK 1971a. Insulin antagonism in pregnancy. In Diabetes Mellitus: Diagnosis and Treatment. SS Fajans and KE Sussman, Co-Editors. New York, American Diabetes Association, Inc. pp. 229-233. Kalkhoff, RK, HJ Kim, J Cerletty, and CA Ferrou 1971b. Metabolic effects of weight loss in obese subjects. Changes in plasma substrate levels, insulin and growth hormone responses. Diabetes 20:83-91. Karam, JH, and GM Grodsky 1963. Excessive insulin response to glucose in obese subjects as measured by immunochemical assay. Diabetes 12:197-204. Keys, A, JT Anderson, and J Brozek 1955. Weight gain from simple overeating. Metabolism 4:427-432. Kinnar, TWG 1963. The pattern of diabetes mellitus in a Nigerian teaching hospital. E Afric Med J 40:288-294. 63. 64. 65, 66. 67. 68. 69. 70. 71. 72. 75. 74. 75. 76. 77. 73. 79. 80. 81. 82. 83. 84. Environmental Factors 103 Labbé, M, and R Boulin 1925. Troubles de la glycorégulacion au cours des infections. Bull et men Soc Med d hosp de Paris 49:1358-1368. Lehmann, WL 1873. Het Arsenigziuer als Geneesmeddel bij Diabetes Mellitus. Amsterdam. Lipman, RL, P Raskin, T Love, J Triebwasser, FR Lecocq, and JJ Schnure 1972. Glucose intoler- ance during decreased physical activity in man. Diabetes 21:101-107. Lundbaek, K 1947. Metabolic abnormalities in starvation diabetes. Yale J Biology and Medicine 20:533-544. Mahler, R 1972. The relationship between eating and obesity. Acta Diabet Lat 9:449-465. Malherbe, C, KC Burrill, SR Levin, JH Karam, and PH Forsham 1972. Effect of diphenylhydantoin on insulin secretion in man. New Engl J Med 286:339-342. Mann, GV, K Teel, O Hayes, A McNalley, and D Bruno 1955. Exercise in the disposition of dietary calories. New Engl J Med 253:349-355. Marks, HH, LP Krall, and P White 1971. Epidemiology and detection of diabetes. Chapter 2 in Joslin's Diabetes Mellitus. A Marble, P White, RF Bradley and LP Krall, Editors. Philadelphia, Lea and Febiger. Melin, AK, and B Ursing 1958. Diabetes mellitus: som komplikation till parotitis epidemica. Nord Medicin 27:1715-1717. Newburgh, LH, and JW Conn 1939. A new interpretation of hyperglycemia in obese middle-aged persons. J Am Med Assoc 112:7-11. Ogilvie, RF 1935. Sugar tolerance in obese subjects. A review of sixty-five cases. Quart J Med 4:345-358. O'Sullivan, JB 1961. Gestational diabetes: Unsuspected, asymptomatic diabetes in pregnancy. New Engl J Med 264:1082-1085. O'Sullivan, JB, and CM Mahan 1964. Criteria for the oral glucose tolerance in pregnancy. Diabetes 13:278-285. Pedersen, J 1954. Weight and length at birth of infants of diabetic mothers. Acta Endocr 34:287-298. Perley, MJ, and DM Kipnis 1967. Plasma insulin responses to oral and intravenous glucose. Studies in normal and diabetic subjects. J Clin Invest 46:1954-1962. Poffenbarger, PL, WL Chick, RL Lavine, S Soeldner, and JH Flewelling 1971. Insulin biosyn- thesis in experimental hereditary diabetes. Diabetes 20:677-685. Porte, D, Jr 1969. Sympathetic regulation of insulin secretion. Its relation to diabetes mellitus. Arch Intern Med 123:252-260. Posner, BI 1973. Insulin metabolizing enzyme activities in human placental tissue. Diabetes 22:552-563. Pyke, DA 1956. Parity and the incidence of diabetes. Lancet 1:818-820. Rabinowitz, D and KL Zierler 1962. Forearm metabolism in obesity and its response to intra- arterial insulin. Characterization of insulin resistance and evidence for adaptive hyper- insulinism. J Clin Invest 41:2173-2181. Rapoport, MI, and HF Hurd 1964. Thiazide-induced glucose intolerance treated with potassium. Arch Intern Med 113:405-411. Rayfield, EJ, RT Curnow, and WR Beisel 1973. Glucose intolerance with paradoxical growth’ hormone secretion during Sandfly fever in man. Clin Res 21:46(a). 104 Diabetes Mellitus 85. 86. 87. 88. 89. 90. 91. 02. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. Rerup, CC 1970. Drugs producing diabetes through damage of the insulin secreting cells. Pharmacol Rev 22:485-520. Rocha, DM, F Santeusanio, GR Faloona, and RH Unger 1973. Abnormal pancreatic alpha cell function in bacterial infection. New Engl J Med 288:700-703. Russell, GFM, and JT Bruce 1964. Capillary-venous glucose differences in patients with dis- orders of appetite. Clin Sci 26:157-163. Salans, LB, JL Knittle, and J Hirsch 1968. The role of. adipose cell size and adipose tissue insulin sensitivity in the carbohydrate intolerance of human obesity. J Clin Invest 47:153-165. Samaan, NA, WA McRoberts, JP Smith, and LG Myers 1971. Metabolic changes in women with tro- phoblastic disease and with intrauterine fetal death compared with metabolic changes during normal pregnancy. J Clin Endocrin 33:521-529. Schless, GL, and GG Duncan 1966. The beneficial effect of intermittent total fasts on the glucose tolerance in obese diabetic patients. Metabolism 15:98-102. Seltzer, HS, and EW Allen 1969. Hyperglycemia and inhibition of insulin secretion during the administration of diazoxide and trichlormethiazide in man. Diabetes 18:19-28, Shambaugh, GE, III, and WR Beisel 1967. Insulin response during tularemia in man. Diabetes 16:369-376. Sims, EAH and ES Horton 1968a. Endocrine and metabolic adaptation to obesity and starvation Am J Clin Nutr 21:1455-1470. Sims, EAH, RF Goldman, CM Gluck, ES Horton, PC Kelleher, and DW Rowe 1968b. Experimental obesity in man. Trans Assoc Am Physicians 81:153-170. Sims, EAH, E Danforth, Jr, ES Horton, JA Glennon, GA Bray, and LB Salans 1972. Experimental obesity in man. A progress report. Israel J Med Sci 8:813-814. Slone, D, LS Taitz, and GS Gilchrist 1961. Aspects of carbohydrate metabolism and kwashiorkor. Brit Med J i1:32-34, Spellacy, WN, FC Goetz, BZ Greenberg, and KL Schoeller 1965. Tolbutamide response in normal pregnancy. J Clin Endoc 25:1251-1254, Spellacy, WN 1969. A review of carbohydrate metabolism and oral contraceptives. Am J Obstet Gynec 104:448-460. Smith, M, and R Levine 1964. Obesity and diabetes. Med Clin N Amer 48:1387-1397. Sussman, KE 1966. Effect of prolonged fasting on glucose and insulin metabolism in exogenous obesity. Arch Int Med 117:343-347. Sweeney, JS 1927. Dietary factors that influence the dextrose tolerance test. Arch Int Med 40:818-830. ‘Talaat, M, YA Habib, AM Higazy, S Abdel Nabry, AY Malek, and ZA Ibraham 1965. Effect of sex hormones on the carbohydrate metabolism in normal and diabetic women. Arch Int Pharmacodyn Ther 154:402-411, Telib, M, S Raptis, KE Schroder, and EF Pfeiffer 1968. Serotonin and insulin release in vitro. Diabetologia 4:253-256. Tzagournis, M, and TG Skillman 1970. Glucose intolerance mechanism after starvation. Meta- bolism 19:170-178. Unger, RH, AM Eisentraut, and LL Madison 1963. The effects of total starvation upon the levels of circulating glucagon and insulin in man. J Clin Invest 42:1031-1039. 106. 107. 108. 109. 110. 111. 112, 113. 114. 115. Environmental Factors 105 Uram, JA, L Friedman, and OL Kline 1958. Influence of diet on glucose tolerance. Am J Physiology 192:521-524. Warren, S, PM LeCompte, and MA Legg 1966. The Pathology of Diabetes Mellitus. Philadelphia, Lea and Febiger. p 107. West, KM, JM Kalbfleisch 1971. Influence of nutritional factors on the prevalence of dia- betes. Diabetes 20:99-108. Whitecar, JP, Jr, GP Bodey, CS Hill, Jr, and NA Samaan 1970. Effect of L-asparaginase on carbohydrate metabolism. Metabolism 19:581-586. Wilkerson, HLC, FK Butler, and J O'S Francis. 1960. The effect of prior carbohydrate intake on the oral glucose tolerance test. Diabetes 9:386-391. Williams, JL, and GF Dick 1932. Decreased dextrose tolerance in acute infectious diseases. Arch Int Med 50:801-818. Yalow, RS, SM Glick, J Roth, and SA Berson 1965. Plasma insulin and growth hormone levels in obesity and diabetes. Ann New York Acad Sci 131:357-371. Young, CM, DL Frankel, SS Scanlon, V Simko, and L Lutwak 1971. Frequency of feeding, weight reduction and nutrient utilization. J Am Diet Assoc 59:473-479. Yudkin, J 1957. Diet and coronary thrombosis. Hypothesis and fact. Lancet 2:155-162. Yudkin, J 1963. Nutrition and palatability. Lancet 1:1335-1338. THE ROLE OF THE NEUROENDOCRINE SYSTEM IN THE DEVELOPMENT OF DIABETES MELLITUS Daniel Porte, Jr. INTRODUCTION Diabetes mellitus is a syndrome in which carbohydrate and lipid fuel transport is grossly disordered. The nature of the primary defect is not understood. However, the resulting hyper- glycemia has been used as an index of a presumed underlying genetic defect. The lack of a more specific marker for the presence or absence of the disease has led to the question of whether all hyperglycemia is due to "diabetes" or not. To answer this question one must consider the role of the neuroendocrine system as an important controller of body glucose metabolism and its role in the development of the diabetic state. A malfunction of this system can produce hyperglycemia. The question then is whether hyperglycemia is due to a neuroendocrine abnormality, to an inter- action of the neuroendocrine system with some other primary factor (''diabetes'), or to "diabetes," independent of the neuroendocrine system. Although hormones have at times been considered to be independent of the central nervous system, it is now apparent that hormonal signals and neural signals are two parts of one inte- grated system for regulation of body metabolism. These factors are inter-organ controllers which are superimposed upon the primary substrate controllers. For example, the concentration of glucose is a primary regulator of its rate of uptake and metabolism by tissues, but this rate is modulated by the neuroendocrine system via several hormones and neural transmitters. To complete the feed- back loop, the output of these neuroendocrine controllers is regulated by the primary substrates, particularly glucose. Finally, these neuro-humoral factors are regulated by each other. The net result of these multiple interactions is to keep the blood glucose within a relatively narrow range. This neuroendocrine control system is no different from any other control system in that the sta- bility of the product (glucose), depends on the multiplicity of the interactions. Therefore a. malfunction of any component part can cause problems in glucose regulation. The purpose of this section is to: 1) describe in simplified terms the complex neuroendocrine regulation of fuels; 2) indicate those areas in which disordered control mechanisms may lead to hyperglycemia; and 3) consider the question of how the hyperglycemia observed clinically may be related to "diabetes' and/or a neuroendocrine abnormality. PHYSIOLOGY OF THE NEUROENDOCRINE SYSTEM The presumed purpose of such a control system is to provide for messages to be given between the organs which have to do with the provision, utilization, and storage of fuels (9). The organs involved include the gastrointestinal tract, which acts as an input source for glucose, amino acids, and fats, and is a source of gut hormones which signal the pancreas, liver, and adipose tissue of the presence of nutrients. Absorbed water soluble nutrients pass first to the liver which acts as an ultimate source for glucose between meals and which serves as a short-term storage system for recently absorbed carbohydrate. The uptake of these nutrients by the liver is regulated as a primary function of their concentration and is secondarily controlled by the hormones insulin, glucagon, 106 The Neuroendocrine System 107 cortisol, and epinephrine, and the autonomic nervous system. The other major fuel storage and re- lease site, adipose tissue is also regulated by the same hormones and nervous factors to provide for a balance between storage of excess calories and provision of fuels between periods of nutrient absorption. Muscle tissue is the primary utilizer of fuels, mostly fatty acids, but does require glucose either as stored glycogen or from plasma glucose for the initiation of exercise. This tissue also acts as a major storehouse of the amino acids which provide for long-term maintenance of blood glucose during starvation after conversion to glucose in the liver and kidney. Thus the liver primarily and the kidney secondarily are the only sources for glucose production during star- vation. All other tissues provide for a constant glucose utilization which is insulin independent. Of these, the central nervous system is the prime user of glucose, metabolizing approximately 80 percent of glucose provided by the liver in the overnight fasted individual. It is the absolute need of the central nervous system for large quantities of glucose which demands the precise and complex regulatory system for maintaining blood glucose levels within a relatively narrow range. The central nervous system therefore is both the major utilizer of glucose and the prime regulatory control center. Although insulin is clearly the major controller of blood glucose concentration, such that any excess or deficiency is associated with major abnormalities in blood glucose regula- tion, this regulation occurs in concert with glucagon, growth hormone, cortisol, epinephrine, sex steroids, and the gut hormones, secretin, pancreozymin, and gastrin. Therefore any change in the output or sensitivity to the effects of any of these other hormones may also be expected to be associated with either hyper- or hypoglycemia. In fact, as discussed later, the concentrations of many of these hormones are abnormal in the diabetic syndrome. The question as to whether or not these hormonal abnormalities are cause or effect will be discussed later. The next section will briefly describe the major metabolic effects of each of these gluco-regulatory hormones. SPECIFIC METABOLIC REGULATORY HORMONES Glucagon. Glucagon is the other pancreatic endocrine hormone. It is an activator of the enzyme adenyl cyclase in many tissues. In the liver this leads to an increase in cyclic AMP which tends to increase glucose output and its administration is therefore followed by hyperglycemia. In adipose tissue glucagon is lipolytic, that is it promotes mobilization of fatty acids. In the pancreas, insulin release is increased, thus modifying the degree of hyperglycemia resulting from glycogen breakdown in the liver. The primary substrate controllers for glucagon are glucose and amino acids. In general, the higher the glucose concentration the lower the glucagon secretion rate, and conversely, the lower the glucose concentration the higher the glucagon secretion rate. Amino acids, on the other hand, stimulate glucagon release and insulin secretion. This bihormonal stimulation maximizes storage of these nutrients in muscle during absorption while preventing simultaneous hypoglycemia. Secondary hormonal controllers for glucagon secretion are epinephrine (27) and pancreozymin, both of which stimulate glucagon release. Neural control of glucagon is via the sympathetic nervous system which stimulates glucagon release when activated (38). A para- sympathetic control has been recently described (7,27). In general terms, therefore, glucagon increases blood glucose and is secreted in response to ingestion of amino acids or activation of the autonomic nervous system by stress or hypoglycemia. Growth hormone. This hormone is secreted by the anterior pituitary gland. Although essential for childhood growth, its primary function in adult man remains to be elucidated. Given in large doses, it may acutely mimic the effects of insulin, but in the physiologic range over a longer 108 Diabetes Mellitus period of time, it opposes the action of insulin on a variety of tissues including muscle, adipose tissue, and the liver. Metabolic control of growth hormone is incompletely understood (39), but glucose clearly appears to suppress its release and hypoglycemia to increase it. Amino acids also stimulate the release of growth hormone as does alpha adrenergic stimulation. The physiologic alpha-adrenergic stimulator may be neurally released norepinephrine or dopamine. Growth hormone is also released during certain phases of sleep. The physiologic significance and control of this phenomenon is unknown. The insulin antagonism that is caused by growth hormone is often overcome by an increase in insulin secretion and thus there may be little effect on blood glucose concen- tration per se. However, should this increase in insulin secretion fail to occur, growth hormone will cause hyperglycemia. Although growth hormone and insulin have been considered antagonistic in glucose homeostasis, this is misleading in the sense that they are synergistic for growth which requires both hormones to be present. Cortisol. The steroid hormone of the adrenal cortex is regulated by ACTH from the anterior pituitary. Cortisol opposes the action of insulin and allows for the effects of several insulin antagonistic hormones to be expressed. These are lipolysis secondary to growth hormone, glucagon and epinephrine in adipose tissue and glycogenolysis and gluconeogenesis secondary to glucagon and epinephrine in the liver. Cortisol also directly increases gluconeogenesis in the liver and thereby tends to increase glucose output by that organ (2). There are no known substrate controllers for the hormone, regulation being provided by pituitary ACTH. Secretion of ACTH itself is controlled by brain centers in the hypothalamus. These centers release peptides called releasing factors which in turn are regulated by cortisol feedback or nerve impulses from the brain's limbic system during periods of stress. Stresses such as hypoglycemia, surgery, etc., that tend to activate the sympathetic nervous system also activate ACTH and growth hormone release. Therefore cortisol, which is insulin antagonistic, is partly responsible for the hyperglycemia of stress (20). Gut hormones--secretion, gastrin, and pancreozymin. These have all been shown to stimulate the secretion of insulin directly and to augment the insulin secretion stimulated by a primary substrate such as glucose or amino acids (15,44,47). Pancreozymin has also been shown to stimulate glucagon secretion. It is presumed that these or other related hormones act as signals from the gastrointes- tinal tract to the pancreas during the process of meal absorption. The presence of food in the stomach is a stimulus for gastrin secretion and the presence of acid in the small bowel is stimulus for secretin, which suggests that these hormones may be relatively nonsubstrate specific. However, one report has claimed that insulin can suppress endogenous secretion levels (10), therefore, it is possible that other endogenous factors control the secretion of these two hormones. Pancreozymin stimulation is believed to be related to a direct stimulatory effect of amino acids in the gastro- intestinal tract. Epinephrine. This catecholamine is released from the adrenal medulla by stress. Epinephrine inhibits the action of insulin in muscle and activates adenyl cyclase in liver and muscle, thereby promoting glycogenolysis in both organs. It also increases gluconeogenesis and from these mecha- nisms leads to an increased glucose release from the liver. Glycogenolysis in muscle is expressed as increased lactate production which after transport from muscle to liver can be used as a sub- strate for gluconeogenesis. In addition, epinephrine directly inhibits insulin secretion by activation of pancreatic adrenergic alpha receptors (43, 45). Epinephrine secretion in turn is controlled by pre-ganglionic sympathetic neurons to the adrenal medulla which release acetyl choline The Neuroendocrine System 109 as a neural transmitter. In a sense then, the adrenal medulla is analagous to a ganglion which is activated as part of the general response to stress. Large doses of glucagon stimulate epin- ephrine release from the adrenal medulla, but a physiological role for this phenomenon has not been described (32). In the pancreas, epinephrine also stimulates the beta cell adrenergic beta receptor to increase beta cell adenyl cyclase. This stimulation of both adrenergic alpha recep- tors which inhibit insulin secretion, and adrenergic beta receptors which stimulate insulin secre- tion provides a push-pull control of the islet beta cell. It prevents complete cessation of islet function even during severe stress, and leads to a super-normal compensatory response to glucose after the termination of the stress-related epinephrine effect (45). Autonomic nervous system. The autonomic nervous system is generally divided into two major, usually antagonistic, sets of neural reflexes which have as effectors either norepinephrine or acetylcholine. Although there may be peripheral dopaminergic and serotinergic neurons, these have not been described well enough to determine whether they play any physiological role as peripheral effectors of the autonomic nervous system. Activation of the sympathetic nervous system produces an integrated response throughout the body leading to changes in all of the organs concerned with fuel metabolism. Stimulation of sympathetic nerves releases norepinephrine which causes glyco- genolysis in liver and mobilization of liver glycogen, lipolysis in adipose tissue with release of free fatty acid, inhibition of insulin secretion from beta cells and stimulation of glucagon release from alpha cells. These are augmented by epinephrine released from the adrenal medulla which in addition to all of these effects increases glycogenolysis in muscle with release of lactic’ acid. Central sympathetic reflexes are also probably involved in the increased secretion of gluco- steroids by activation of ACTH release and increased growth hormone output, possibly through central dopaminergic neurons. Most of this system has been well appreciated for many years, but despite a well described neural inervation of the islets of Langerhans by early investigators, the incorporation of the islets into the neuro-endocrine system is very recent (56). Activation of the parasympathetic nervous system releases acetylcholine which produces effects which are opposite to the metabolic effects of sympathetic stimulation. There is an increase in glucose uptake by the liver, and a decreased glucose output from the liver; an increase in insulin secretion and an increase in the release of gut hormones. Our understanding of substrate feedback to these two systems is really embryonic, but evidence has been presented for central glucose receptors which are insulin dependent (11). These receptors have been related to two hypothalamic nuclei. Electrical stimulation of the ventromedial hypothalamus is associated with a sympathetic-like response with a decrease in plasma insulin and an increase in plasma glucagon (21). This suggests these centers activate nerves to the adrenal, liver, and pancreas. Stimulation of the lateral hypothalamus, on the other hand, has been reported to cause a decrease in blood glucose and an increase in plasma insulin (51). This would suggest this center to activate parasympathetic nerves to the liver and pancreas. The fact that one of the more reliable methods of producing hyperinsulinism and obesity is to destroy the ventromedial hypo- thalamic nucleus illustrates the importance of these nuclei to body metabolism (3,18,19,55). Such a procedure is reliably followed first by increased insulin secretion and decreased growth hormone, then by increased eating and obesity. 110 Diabetes Mellitus Modulation of physiologic adaptations. The neuroendocrine system acts in a concerted fash- ion to provide for a smooth interorgan adaptation to several major physiologic states--fasting, exercise, stress, feeding, and growth. The transition from the fed to the fasted state appears to be an integrated sympathetic-like response with a decrease in insulin secretion, and an increase in glucagon, growth hormone, and epinephrine secretion. Exercise appears to be a more powerful stimulus of the same type in which there is a decreased insulin secretion, increased glucagon, growth hormone and epinephrine, and clear-cut activation of the entire cardiovascular related peripheral sympathetic nervous system. Finally, in the most severe form of stress, such as severe exhaustive exercise or a variety of pathophysiologic events, such as burns, myocardial infarction, hypothermia, severe infection, shock, and hypoxia, there is an increased output of ACTH and cortisol as well. In general terms then, activation of the sympathetic system produces physiologically nor- mal but diabetic-like states in non-diabetic persons (25,38,43,44,45), Feeding appears to be an example of an integrated parasympathetic response, but the overall pattern of response is dependent upon the nutrient absorbed. Glucose ingestion results in an output of gut hormones, particularly secretin, followed by an increase in insulin, a decrease in glucagon, a decrease in growth hormone, and possible a reduction in ACTH (46), although this latter phenomenon is uncertain at the present time. There is also some suspicion that there is a simul- taneous suppression of sympathetic nervous system activity, but this has not been documented except for the observation that glucose administered directly into the central nervous system will alter firing of some hypothalamic neurons (1, 56). The pattern of response after amino acid ingestion is a similar parasympathetically oriented response except for an increase in glucagon and an increase in growth hormone which appears to be directly substrate mediated. Growth, or the long-term adaptation to nutrient ingestion appears to be similar to the feeding response. The nature of the neuroendocrine response has not been well characterized, but growth hormone, thyroid hormone, and insulin are all necessary for growth to occur. METABOLIC PATHOPHYSIOLOGY OF THE NEUROENDOCRINE SYSTEM Just as deficient insulin secretion will produce hyperglycemia, so will oversecretion of any of the insulin antagonistic hormones, glucagon, cortisol, growth hormone, and epinephrine, or activation of the sympathetic nervous system. The clinician may be hard pressed at times to differ- entiate the hyperglycemia produced under these conditions from the genetic diabetic syndrome. In many such instances the patient returns entirely to normal after a period of hyperglycemia. In others some residual disability may be maintained. Which patient has "diabetes"? The situation is further confused by the fact that one patient with a neuroendocrine hormonal abnormality does not have hyperglycemia of any significant degree, whereas another may be severely hyperglycemic and require insulin therapy. Are some patients unusually sensitive to these abnormal hormonal states, or has metabolic stress selected individuals with an inherent genetic diabetic defect which is only expressed when some other pathologic abnormality appears? At the moment there appears to be no simple answer to this question, but to approach it we must first consider the effect of major abnor- malities of each of the glucoregulatory hormones. Excess cortisol secretion. Primary excess cortisol secretion may be due either to a primary adrenal neoplasm with autonomous hormonal release or to a primary pituitary or hypothalamic abnor- mality resulting in excess ACTH secretion and secondary hypercortisolism. In either case an insulin The Neuroendocrine System 111 resistant state is produced which is coupled with an excessive breakdown of body protein and conversion of mobilized amino acids into glucose through the process of gluconeogenesis in the liver (2). In some individuals compensation occurs by increased secretion of insulin while in others this compensation fails and hyperglycemia ensues (12, 44). Iatrogenic hypercortisolism is now extremely common due to the frequent use of these compounds in various rheumatic diseases. The degree and frequency of hyperglycemia appears to depend upon (a) the dose of cortisol used; (b) the duration for which it is given; and (c) the underlying genetic constitution of the indivi- dual (40). Low dose steroids have been used to detect individuals with the presumed diabetic gene. No long-term follow-up of such individuals has been made and therefore it is unknown whether persons developing hyperglycemia after steroid administration have in fact been identified as genetically abnormal people. Since excessive steroid secretion occurs in response to a variety of stresses, it is reasonable to assume that such stress-induced hypercortisolism must contribute to the hyper- glycemia observed. Excess growth hormone. Studies in patients with pituitary tumors secreting excessive amounts of growth hormone (acromegaly) have indicated that there is a general resistance to the hypoglycemic effects of insulin. These patients have increased basal and stimulated insulin levels (44). The mechanism for this resistance is unknown. Many patients with acromegaly will show some deteriora- tion of glucose tolerance and occasionally develop clinically overt diabetes. Some workers believe that overt diabetes will only occur in those subjects with some genetic predisposition (35), but firm evidence for this belief is not yet available. Whenever excess growth hormone is found, an increase in insulin resistance would be expected to contribute to any hyperglycemia observed. Excessive glucagon secretion. Primary hypersecretion of glucagon has only been reported rarely from tumors derived from islet alpha cells. It is a potent glycogenolytic and gluconeo- genic hormone and these properties are evident in such rare patients with a glucagonoma whose diabetic state is difficult to clinically distinguish from other causes of the diabetic syndrome (37). Milder abnormalities of glucagon secretion are common in the diabetic syndrome particularly associated with sympathetic nervous system activation. The etiology of this finding will be discussed later. Excessive secretion of epinephrine and/or norepinephrine. Tumors of the adrenal medulla are regularly associated with hyperglycemia and glucose intolerance. These hormones are known to be glycogenolytic in both muscle and liver, to inhibit insulin secretion, and to produce antagonism to the peripheral effects of insulin. As circulating hormones they may also activate the sympa- thetic nervous system and promote further hyperglycemia via neural reflexes (14). Thyrotoxicosis. Excessive thyroid hormone secretion has been associated with an increased frequency of carbohydrate intolerance. The mechanism of this association is not completely under- stood but it may be related to increased sensitivity of peripheral tissues to the effects of catecholamines. In some instances there has been improvement of carbohydrate tolerance by the use of catecholamine-depleting drugs (44). Long-term follow-up of patients who develop hyperglycemia during thyrotoxicosis has not been done. Therefore it is again unknown whether this is purely a physiologic or pharmacologic effect of thyroid hormone or whether a genetically susceptible indivi- dual has been detected. 112 Diabetes Mellitus Stress states. A variety of stress states have been associated with hyperglycemia and glyco- suria. Thess include trauma, surgery, hypovolemic shock, burns, myocardial infarction, hypothermia, and severe psychologic stress (43, 44, 45). Each of these is characterized by generalized activa- tion of the sympathetic nervous system, with stimulation of epinephrine release from the adrenal medulla as well as high levels of ACTH and cortisol, and growth hormone. More recently hyperglu- cagonemia has also been reported (6,31,38,49). These all appear to be coordinated effects which can be explained by activation of the sympathetic nervous system. In some cases, alpha adrenergic blocking agents or catecholamine depleting drugs have been used to treat the hyperglycemia observed. In other instances, insulin in combination with glucose, has been administered. In both cases improved mortality and morbidity have been reported to result from this treatment (44, 45). There has been no evidence presented to suggest that some or any of these patients have an underlying genetic defect, but it would seem likely that a person with a genetic predisposition to the diabetic syndrome would be more likely to develop hyperglycemia in the presence of severe stress than a gen- etically normal subject. Even the mild stress of simple exercise is also associated with inhibition of insulin secretion, increased levels of glucagon, and accelerated glucose production (16). There- fore, there appears to be a continuum from the physiologic response to exercise to the pathophysi- ologic stress state. The importance of social stress to hyperglycemia remains largely unexplored. Whether such stress related events can be looked upon as beta cell injuries which contribute to the eventual development of permanent islet damage is unknown. POSSIBLE CONTRIBUTIONS OF THE NEUROENDOCRINE SYSTEM TO ACUTE COMPLICATIONS OF DIABETES Ketoacidosis. In this form of severe absolute insulin deficiency, growth hormone, glucagon, and catecholamine levels have all been shown to be elevated. These elevations are partly if not completely reversed by administration of insulin. All of these changes are consistent with acti- vation of the sympathetic nervous system and in many ways the whole syndrome of ketoacidosis suggests a caricature of severe stress. Much of what happens in ketoacidosis can be mimicked by administra- tion of these hormones and neuro-transmitters and therefore the question has been raised as to exactly how much of this syndrome is dependent upon them (43). It has been shown that the ketosis is directly dependent upon the mobilization of free fatty acids and can be reversed by inhibitors of fatty acid mobilization, including those that are believed relatively specific for the sympa- thetic nervous system. It has also been observed that some patients with clinical ketoacidosis may fully recover and not require permanent insulin therapy. This could be explained if the aggra- vation of the diabetic syndrome were in some way related to reversible elevations of epinephrine, norepinephrine, glucagon, growth hormone, and cortisol. Whether or not these factors are truly etiological, there can be no doubt that the severity of the syndrome must in part depend upon their presence. In some instances patients have been treated with adrenergic blocking agents with a reduction in the frequency of ketoacidosis (4). It would appear that the insulin deficient dia- betic is unable to compensate for these counterinsulin hormonal states and in this sense the diabetic state confers an increased susceptibility to the effects of stress on blood sugar. The question then arises as to what role stress and hormonal aberrations play in the day-to-day control of blood sugar in the clinically diagnosed diabetic. Although no well-documented long-term case studies of this effect are available, in short-term controlled studies it has been found that hypnosis and a stressful interview will increase ketone body levels more in a diabetic than in a normal subject (26). The Neuroendocrine System 113 It is also a frequent clinical observation that psycholgical factors play an important role in the day-to-day regulation of glucose level in the diabetic patient. It seems reasonable to hypothesize that it is the neuroendocrine system which is responsible for these effects. CONTRIBUTIONS OF THE NEUROENDOCRINE SYSTEM TO CHRONIC COMPLICATIONS OR CONCOMITANTS OF THE DIABETIC SYNDROME The diabetic syndrome consists of four associated abnormalities: 1) microvascular disease, associated with capillary basement membrane thickening; 2) large vessel arterial disease, particu- larly accelerated atherosclerosis; 3) neuropathy and loss of nerve transmission function; and 4) hyperglycemia and metabolic abnormalities. In general, it is believed that the neuroendocrine system has a role only in the regulation of blood glucose. However, evidence has been produced that this system may play a role in the microvascular disease and the accelerated atherosclerosis. Microvascular disease. The chance observation that a hypophysectomized patient had a spon- taneous remission of diabetic retinopathy has led to considerable effort to determine whether the pituitary produces something which exacerbates diabetic retinopathy. An extensive series of pituitary ablations seems to indicate that there is some amelioration of the progression of the disease and that this occurs despite the replacement of thyroid hormone and cortisol. This has left the gonadotropins, prolactin, and growth hormone as potential candidates for the necessary factor for progression of retinopathy. Most attention has been focused on growth hormone, but the others have not been definitively ruled out. Whether growth hormone secretion is abnormal in diabetes is not totally clear and this will be discussed below. Nevertheless, in the only double-blind series studied, about 50 percent of the patients hypophysectomized have amelioration of the retinopathy (30). Since growth hormone secretion is normally related to a central dopaminergic alpha adrenergic mechanism (39), central adrenergic factors may be related in part to the development of this particu- lar type of vascular degeneration (42). Accelerated atherosclerosis--role of hormones. Accelerated atherosclerosis occurs as a regular feature of the diabetic syndrome. The mechanism of this association is unknown, but may involve a host of complicating interacting factors such as hormonal factors, lipid abnormalities, and obesity. The best evidence for direct hormonal involvement in atherosclerosis has been related to cortisone, thyroid hormone, insulin, and stress. Cortisol. Accelerated atherosclerosis has been associated with hypercortisolism whether in- duced by therapeutic administration of steroids or in patients with pituitary hypersecretion of ACTH (22). Whether this is a direct effect of the steroids on the vessel wall or related to some other induced abnormality, such as the associated hyperinsulinism, or hyperlipidemia, is unclear. Insulin. Hyperinsulinism has been reported in a variety of atherosclerotic subjects when compared with control groups, although in some cases the controls were not well matched (53). Such an association is of interest because there is now evidence that the arterial smooth muscle cell is insulin sensitive (52). It is well known that insulin is a lipogenic hormone. One hypothesis has been that excessive amounts of insulin or high levels of insulin at inappropriate intervals may predispose this insulin sensitive cell to atherosclerosis (53). Although it is clear that in dia- betes there is relative deficiency of insulin secretion in response to challenge, the frequently associated obesity is itself associated with excess insulin secretion in the basal state. There- fore, on an absolute basis, many diabetic subjects have basal hyperinsulinism even if there is simultaneous relative deficiency of the insulin response. The exposure to high levels of insulin 114 Diabetes Mellitus may be exacerbated by the deficient early insulin response which then leads to a subsequent pro- longation of the insulin response to meal challenge. In this sense insulin levels are often elevated at inappropriate periods. Even the insulin treated patient may have this problem, since therapy with insulin involves the administration of long-acting preparations in quantities which may be superphysiologic either due to the fact that there are antibodies to insulin or because the normal intermittent secretion pattern of insulin cannot be mimicked by injection therapy, resulting in inappropriately high insulin levels during part of the 24 hour period. Thyroid. Hypothyroidism has been used as an experimental tool to produce plasma lipid abnor- malities and accelerated atherosclerosis for many years (22). In some species this is the only means to induce atherosclerosis, therefore it is not surprising that clinical hypothyroidism in man has been associated with accelerated atherosclerosis. Here again, the clinical studies are not as carefully controlled as the experimental animal. Recently the squirrel monkey has been compared on an atherogenic diet with and without hypothyroidism and there was a striking difference in re- spect to coronary atherosclerosis (33). Therefore the question has been raised as to whether or not subclinical hypothyroidism may contribute to the atherosclerosis problem. Catecholamines and stress. Subjects with an increased risk for coronary heart disease have been identified by psychological testing (17). It has been suggested that their biologic response to life stress is the mechanism by which personality relates to atherosclerosis. The pathogenic link between life stress situations and the development of atherosclerosis needs to be documented and explored, but almost certainly involves activation of the sympathetic nervous system, with elevated levels of cortisol, glucagon, catecholamines, growth hormone, and inappropriate secretion of insulin. It is of interest that there is an association between the degree of hyperglycemia during glucose loading and the presence of atherosclerotic complcations (13). Since activation of the sympathetic nervous system would be expected under stress conditions, one could postulate that either the hyperglycemia is associated with coronary artery disease because of stress, or that stress creates changes in metabolic regulation which themselves induce accelerated atherosclerosis. Some have hypothesized that fuels have been mobilized as a response to stress for expected muscular work, but in the presence of a sedentary existence these fuels are rechanneled into lipoprotein production leading to higher lipid levels than expected (54) and accelerated atherosclerosis. Although the diabetic would not be expected to have any unusual propensities for this relationship, diabetes mellitus resembles stress in many ways, and it therefore seems possible that the same series of hormones contribute to atherosclerosis in the diabetic by a mechanism which has much in common with sustained stress. In addition, stress in a diabetic may be expected to be less well counter-regulated than in otherwise metabolically normal persons. EVIDENCE FOR A ROLE OF THE NEUROENDOCRINE SYSTEM AS A PRIMARY ETIOLOGIC AGENT IN THE DIABETIC SYNDROME As discussed above, it is not unusual to find elevated concentrations of growth hormone, catecholamines, glucagon, and cortisol in subjects without obvious stress who have been diagnosed as having diabetes. Although in many instances it is possible to relate these abnormalities to either an associated stressful event, or to the stress of one of the complications or concomitants of the diabetic syndrome, some of the same disorders of hormonal regulation have been reported in patients with the mildest abnormalities of carbohydrate tolerance. For example, in a group of The Neuroendocrine System 115 "prediabetic subjects,' those with two known diabetic parents and a normal glucose tolerance test, an inappropriately high growth hormone response to glucose was noted during the test which was not present in normal subjects (50). Similarly, plasma glucagon levels have been found to be either abnormally elevated in the basal state or to lack the normal suppression after glucose loading; (8,23,41). There have been few if any studies of catecholamines or cortisol in such subjects, but there is one patient reported by Robertson who was found to have an apparently overactive alpha-adrenergic receptor (48). Since the central nervous system is known to contain glucose-responsive neurons which are in or near the hypothalamus, one is tempted to integrate these findings and suggest that all of the hormones known to be related to the sympathetic nervous system may be inappropriately activated. Whether this activation represents some primary phenomenon in the diabetic syndrome or is secondary to a basic abnormality of the pancreas is unknown but there is no a priori reason for sug- gesting one possibility over the other. There is a considerable body of evidence that there is a basic problem in the recognition of glucose by the beta cells of the endocrine pancreas. It seems conceivable that this problem may be more generalized than realized and that glucose may not be adequately recognized in other endocrine cells, such as the alpha cell, or in central regulators of fuel metabolism, such as the glucose receptors of the hypothalamus. Certainly the levels of many counter regulatory hormones appear to be inappropriate for the hyperglycemia found in diabetes because they are normal or elevated at a time when they might be expected to be suppressed. One other potential role for the neuroendocrine system as an etiologic agent in diabetes is in relation to the observation that acute injuries often appear to be in some way related to the on- set of diabetes mellitus. Although genetic constitution obviously plays an important part in the underlying etiology of the disease, it has been suggested that the apparently low penetrance may be due to the need for some additional factor to precipatate the clinical illness. Infection, surgi- cal procedures, myocardial infarction, etc. have often seemed to play an important role in this precipitation. The question arises as to whether repeated abnormalities in neuroendocrine control might provide a form of injury which eventually leads to clinical diabetes mellitus. It is of some interest that in a study by Loubatieres (34) of the effects of catecholamines on pancreatic endocrine function in dogs, partial pancreatectomy followed by infusions of catecholamines appeared to result in a permanent diabetic syndrome. Thus the neuroendocrine system, particularly in terms of stress activation, may be one of the injurous factors which precipitates the expression of the clinically overt diabetic state. RELATION OF THE NEUROENDOCRINE SYSTEM TO INSULIN RESISTANCE Ever since the demonstration that hypophysectomy or adrenalectomy markedly increase peripheral sensitivity to the effects of an injected dose of insulin, it has become clear that hormones play a major role in determining the sensitivity of many tissues to insulin. Hormones, such as cortisol, growth hormone, glucagon, and epinephrine all induce insulin resistance and have already been dis- cussed. There are several clinical syndromes such as uremia, pregnancy, and chronic liver disease where resistance to the action of insulin has also been demonstrated. In some cases it may be that hormones are playing a role in the nature and degree of this resistance. However in obesity, the most common and important cause of insulin resistance, there is no evidence that this resistance is related to hormonal factors. Growth hormone, in fact, is usually believed to be lower than 116 Diabetes Mellitus normal, and although there is increased secretion rate of corticosteroids, plasma levels are normal (5). On the other hand, the genesis of the obese state probably involves some disturbance of the hypothalamic recognition of total fat mass. There is a series of studies indicating that in some animal models with spontaneous obesity the weight gain is regulated (55). That is, if additional calories are force-fed, there will be a return to the original overweight state when forced over- feeding is stopped. Similarly, if calories are restricted, weight is lost, only to be regained to the original state (but not beyond) when free feeding is allowed. This phenomenon is common in human obesity. Obese subjects rarely have difficulty in losing weight for a short period of time, but often regain it promptly. This has led to the concept that human obesity may also be a regulated obesity (55). To explain this phenomenon one should consider the ventromedial hypotha- lamic lesion as one of the striking animal models for obesity with a known pathogenesis (55,56). When obesity is produced by destruction of the ventromedial nucleus of the hypothalamus these animals immediately become hyperinsulinemic which is followed by overeating and cbesity. If they are force-fed after reaching a new stable elevated weight, they will immediately lose any excess weight due to the force feeding and return to their original degree of obesity. In this case then, damage to the hypothalamus leads to a regulated obesity. It has usually been assumed that the association between diabetes and obesity in man is due to the stress that the insulin resistance of obesity places upon the islet, requiring increased secretion of insulin, both in the basal and stimulated states. However, in the ventromedial lesioned animals, it is quite clear that there is a central defect which leads to hyperinsulinism prior to the obesity, and the insulin resistance then follows. Although at this time there is no solid evidence for it, it is possible that the clinical associations of these two syndromes may in certain instances be related to some central defect in appetite regulation in the genetically diabetic subject which leads first to hyper- insulinism and obesity and only later to defective insulin release and diabetes. KNOWLEDGE GAPS It is obvious from this discussion that the neuroendocrine system and its controls are only understood in an elementary way at the present time. There is a particular lack of knowledge of the central integrating mechanisms for the neuroendocrine system and how communication between the periphery (e.g., the size of the adipose mass) and the appetite regulating centers occurs. Further- more, the nature of the central receptors and their location, sensitivity, and to which specific hormones they are related are only poorly understood. It should also be clear that the action of any one hormone may be influenced and altered by the simultaneous action of another, and that endocrine actions often include stimulatory as well as repressive activities upon the secretion of the glands. Although these are qualitatively describable at the present time, the quantitative nature of these interactions is very incomplete. Therefore the relative importance of substrate controllers versus hormonal controllers to the overall regulation of fuel metabolism cannot as yet be assigned with any precision. The finding that there are many hormonal aberrations in diabetes mellitus in addition to defective insulin secretion still leaves the question of whether these hormonal aberrations are secondary to the underlying inability to secrete insulin or whether some of these hormonal aberrations produce the diabetic syndrome or influence insulin secretion per se as a primary event. Perhaps one of the most important unknowns is the nature of the controller The Neuroendocrine System 117 for integration of appetite and weight. There seems to be no doubt that the frequency of clinical diabetes mellitus is very definitely increased in populations of excess body weight. If this fac- tor could be understood and treated one might expect a marked amelioration of at least the carbo- hydrate abnormalities that have been observed. Whether chemical or other modifications of this system in an attempt to restore the abnormal hormonal set of the diabetic to a more normal physi- ologic function is useful or feasible is totally unknown, and cannot even be approached at the present time without further understanding of the basic system. It would seem that knowledge at the present moment is poised at the point where there is general recognition that hormones, such as insulin, are not solely regulated by primary substrate messages delivered through the circula- tion, but are also importantly controlled by circulating hormone controllers and by direct neural input as well. The integration of the neuroendocrine system through the autonomic nervous system is at best imperfectly understood and is the major factor preventing the determination of its influence upon the diabetic state. In order to rectify these gaps, it will be necessary to have an integrated interaction between psychologists, neuroendocrinologists, and clinicians interested in the problem of diabetes melli- tus. At the present time, people in each of these disciplines are interested in their own subset of problems relating to the overall system, but there has been very little interaction between any of the three, except for rare collaboration between investigators with neurophysiological and clinical experience, or psychologists with persons experienced in metabolic problems. It would seem at the present time that if major progress were to be realized, it would necessarily require the collaborative efforts of these disciplines. A major gap in clinical information relates to the long-term follow-up of individuals who have been identified during some period of stress as having an abnormal carbohydrate metabolism during stress. That is, in subjects who develop hyperglycemia in the presence of myocardial infarction, surgery, infection, or administration of estrogens, steroids, etc., there has been almost no information available as to what percentage of these subjects eventually develop clinical diabetes mellitus as compared with appropriate control groups. Although it is commonly assumed that such stresses have in fact identified genetically abnormal persons, only a long-term cooperative study of such individuals could realize this potential. This is of course important if we are to under- stand the natural history of the diabetic syndrome and would be necessary if we are to identify those individuals in whom preventive measures should be undertaken. REFERENCES 1. Anand, BK, GS China, and B Singh 1962. Effect of glucose on the activity of hypothalamic "feeding centers.'" Science 138:597-598. 2. Ashmore, J, and G Weber 1968. Carbohydrate Metabolism and Its Disorders. Vol II. Academic Press, New York, pp 336-374. 3. Baile, CA, and MJ Forbes 1974. Control of feed intake and regulation of energy balance in ruminants. Phys Rev 54:160-214. 4. Baker, L, A Barcai, R Kaye, and N Haque 1969. Beta adrenergic blockade and juvenile diabetes: Acute studies and long term therapeutic trial. J Peds 75:19-29. 5. Bjorntorp, P 1972. Disturbances in the regulation of food intake. Obesity: anatomic and physiologic-biochemical observations. Adv Psychosom Med 7:116-147. 118 Diabetes Mellitus 10. 11. 12. 13, 14. 15. 16. 17. 18. 19. 20. 21. 22. 23, 24. 25. 26. 27. Bloom, SR, PM Daniel, DI Johnston, O Ogawa, and OE Pratt 1973. Release of glucagon, induced by stress. Quart J Exp Phys 58:99-108. Bloom, SR, AV Edwards, and NJA Vaughn 1974. The role of the autonomic innervation in the control of glucagon release during hypoglycemia in the calf. J Physiol 236:611-623. Buchanan, KD, and AM McCarroll 1972. Abnormalities of glucagon metabolism in untreated diabetes mellitus. Lancet 2:1394-95. Cahill, GF, and OE Owen 1968. Carbohydrate Metabolism and Its Disorders. Vol I. London, Academic Press pp 497-522. Chisholm, DJ, I Lazarus, JD Young, and E Kraegen 1970. Abnormalities of serum secretion in diabetes mellitus. Diabetes 19:365. Debons, AF, I Krimsky, and From 1970. A direct action of insulin on the hypothalamic satiety center. Am J Physiol 219:938-943. Ensinck, JW, and RH Williams 1972. Handbook of Physiology. Section 7: Endocrinology. Vol I. The Endocrine Pancreas. Washington, DC, American Physiological Society. Chap 43. Epstein, FH 1967. Hyperglycemia, a risk factor in coronary heart disease. Circ 36:609. Ezdinli, EZ, R Javid, G Owens, and JE Sokal 1968. Effect of high spinal cord section on epinephrine hyperglycemia. Am J Physiology 214:1019-1024. Fajans, SS, and JC Floyd, Jr 1972. Handbook of Physiology. Section 7: Endocrinology. Vol I. The Endocrine Pancreas. Washington, DC, American Physiological Society. Chap 30. Felig, P, J Wahren, R Hendler, and G Ahlborg 1973. Plasma glucagon levels in exercising man. N Engl J Med 281:184. Friedman, M, SO Byers, RH Roseman, and FR Elevitch 1970. Coronary-prone individuals (type A behavior pattern). Some biochemical characteristics. JAMA 212:1030-1037. Frohman, LA 1969. The endocrine function of the panceras. Ann Rev Physiol 31:353-382. Frohman, LA 1971. Diabetes Mellitus. Diagnosis and Treatment. Vol III New York, American Diabetes Association. Chap XI. Frohman, LA 1972. Clinical neuropharmacology of hypothalamic releasing factors. New Engl J Med 286:1391-1397. Frohman, LA, and LL Bernardis 1971. Effect of hypothalamic stimulation on plasma glucose, insulin, and glucagon levels. Am J Physiol 221:1596-1603. Furman, RH 1969. Atherosclerosis; Pathology, Physiology, Aetiology, Diagnosis and Clinical Management. Amsterdam, London, New York, Elsevier. Chap 6. Gerich, JE, M Langlois, C Noacco, JH Karam, PH Forsham 1973. Lack of glucagon response to hypoglycemia in diabetes: Evidence for an intrinsic pancreatic alpha cell defect. Science 182:171-173. Goodner, CJ 1971. Diabetes Mellitus. Diagnosis and Treatment. Vol III. New York, American Diabetes Association. Chap 1. Goodner, CJ, and D Porte, Jr 1972. Handbook of Physiology. Section 7: Endocrinology. Vol I. The Endocrine Pancreas. Washington, DC, American Physiological Society. Chap 38. Hinkle, LE, GB Conger, and S Wolf 1950. Studies on diabetes mellitus: The relation of stressful life situations to the concentration of ketone bodies in the blood of diabetic and nondiabetic humans. J Clin Invest 29:754-769. Iversen, J 1973. Adrenergic receptors and the secretion of glucagon and insulin from the isolated, perfused canine pancreas. J Clin Invest 52:2102-2116. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42, 43. 44. 45. 46. 47. 48. The Neuroendocrine System 119 Iversen, J 1973, Effect of acetylcholine on the secretion of glucagon and insulin from the isolated, perfused canine pancreas. Diabetes 22:381-387. Lacy, PE, and MH Greider 1972. Handbook of Physiology. Section 7: Endocrinology. Vol I. The Endocrine Pancreas. Washington, DC, American Physiological Society. Lancet Editorial 1969. Pituitary destruction for diabetic retinopathy. Lancet 2:415-417. Lanido, S, P Segal, and B Esrig 1973. Secretion of endogenous immuno-reactive glucagon following acute myocardial infarction in man: its role in the pathogenesis of post-infarction hyperglycemia. Am J Cardiol 31:144. Lawrence, AM 1967. Glucagon provocative test for pheochromocytoma. Ann Intern Med 66:1091-1096. Lehner, .NDM, TB Clarkson, and HB Lofland 1971. The effect of insulin deficiency, hypothy- roidism and hypertension on atherosclerosis in the squirrel monkey. Exp Molec Path 15:230-244. Loubatieres, A, MM Mariani, J Chapel, J Taylor, MH Houareau, and AM Rondot 1965. Action nocive de 1'adrenaline pour la structure histologique des ilots de langerdans du pancreas. Action protectrice de la dihydroergotamine. Diabetologia 1:13. Luft, R, E Cerasi, and CA Hamberger 1967. Studies on the pathogenesis of diabetes in agromegaly. Acta Endocr 56:593-607. Malaisse, WJ 1972. Handbook of Physiology. Section 7: Endocrinology. Vol I. The Endocrine Pancreas. Washington, DC, American Physiological Society. Chap 4. Mallinson, CN, SR Bloom, AP Warin, PR Salmon, and B Cox 1974. A glucagonoma syndrome. Lancet 2:1-5. Marliss, EB, L Girardier, J Seydoux, CB Wolheim, Y Kanazawa, L Orci, AF Renold, and D Porte, Jr 1973. Glucagon release induced by pancreatic nerve stimulation in the dog. J Clin Invest 52:1246-1260. Martin, JB 1973. Neural regulation of growth hormone secretion. New Engl J Med 288:1384-1393. McKiddie, MT, M Jasani, KD Buchanan, JA Boyle, and WW Buchanan 1968. The relationship between glucose tolerance plasma insulin, and corticosteroid therapy in patients with rheumatoid arthritis. Metab Clin Exptl 17:730-740. Miller, WA, GR Faloona, E Aguilar-Parada, and RH Unger 1970. Abnormal alpha-cell function in diabetes. New Engl J Med 283:109-115. Passa, P, C Gauville, and J Canivet 1974. Influence of muscular exercise on plasma level of growth hormone in diabetics with and without retinopathy. Lancet 2:72-74. Porte, D, Jr 1969. Sympathetic regulations of insulin secretion. Arch Intern Med 123:252-260. Porte, D, Jr, and JD Bagdade 1970. Human insulin secretion: an integrated approach. Ann Rev Med 21:219-240. Porte, D, Jr, and RP Robertson 1973. Control of insulin secretion by catecholamines, stress, and the sympathetic nervous system. Fed Proc 32:1792-1796. Rayyis, SS, and JE Bethune 1970. Effect of blood glucose on ACTH secretion in man. Clin Res 18:171. Rehfeld, JF 1972. Gastrointestinal hormones and insulin secretion. Scand J Gastroenterology 1:289-292. Robertson, RP, JD Brunzell, WR Hazzard, RL Lerner, and D Porte, Jr 1972. Paradoxical hypo- insulinemia: an alpha adrenergic mediated response to glucose. Lancet 2:787-789. 120 Diabetes Mellitus 49. 50. 51. 52, 5%. 54, 55. 56. Rocha, DM, MF Santevsanio, GR Faloona, and RH Unger 1973. Abnormal pancreatic alpha-cell function in bacterial infections. N Engl J Med 288:700. Sonksen, PH, JS Soeldner, RE Gleason, and G Boden 1973, Abnormal serum growth hormone re- sponses in genetically potential diabetic male patients with normal oral glucose tolerance: Evidence for an insulin-like action of growth hormone in vivo. Diabetologia 9:426-437. $s Steffens, AB, GJ Mogenson, and JAF Stevenson 1972. Blood glucose, insulin, and free fatty acids after stimulation and lesions of the hypothalamus. Am J Physiol 22:1446-1452, Stout, RW, R Ross, and EL Bierman 1973. The arterial smooth muscle cell; the effect of insulin on cell proliferation. Europ J Clin Invest 3:2171. Stout, RW, and J Vallance-Owen 1969. Insulin and atheroma. Lancet 1:1078. Taggart, P, and M Carruthers 1971. Endogenous hyperlipidaemia induced by emotional stress of racing driving. Lancet 1:363. Woods, SC, E Decke, and JR Vasselli 1974. Metabolic hormones and regulation of body weight. Psych Rev 81:26-43. Woods, SC, and D Porte, Jr 1974. Neural control of the endocrine pancreas. Physiol Rev 54:596-619. Chapter Chapter Chapter Chapter 10 11 12 13 ACUTE COMPLICATIONS OF DIABETIC STATE Diabetic Coma Daniel W. Foster Hyperosmolar Coma James B. Field Lactic Acidosis: Interrelationships with Diabetes Mellitus and Phenformin Robert A. Kreisberg Acute Complications of the Diabetic State Joseph Silva and F. Robert Feckety 123 133 142 154 10 DIABETIC COMA Daniel W. Foster BACKGROUND The diabetic patient is vulnerable to a series of acute metabolic complications. As indicated in Figure 1, the insulin ‘deficiency which characterizes this disease may result in a spectrum of disorders which tend to cluster in three recognizable (but overlapping) syndromes : (1) Relatively pure ketoacidosis (minimal or no dehydration), (2) Ketoacidosis with dehydration, (3) Relatively pure dehydration (minimal or no ketosis). Ketoacidosis H I . Insulin NORMAL Insulin Kersoridasis 0 cemia ypogly Excess METABOLISM Deficiency Deh mtion Dehydration The first two syndromes are commonly designated as ''diabetic coma' or "diabetic ketoacidosis," while the third is generally defined as hyperosmolar, nonketotic coma. From a physiologic standpoint, the clinical presentations can be understood by considering that uncontrolled diabetes is manifested by severe hyperglycemia, which causes dehydration, and by overproduction of ketone bodies in the liver, which results in a metabolic acidosis. If the former predominates, the clinical picture is that of hyperosmolar coma, to be discussed in a subsequent section, while if the latter takes precedence, the presentation will be that of diabetic coma, the subject of this discussion. It should be emphasized in the beginning that diabetic coma is not a rare condition, al- though patients at risk are primarily those with severe enough disease to require insulin therapy. At the Los Angeles County-University of California Medical Center, there were 340 admissions for ketoacidosis (with 32 deaths) in a 3-year period (6). Since each episode may be life threatening to the individual and since each admission requires a significant period of hospitalization, it is clear that advances in the understanding of the ketogenic process, which might lead to prevention or more rapid treatment, would be of substantial benefit to the individual diabetic and to society as a whole in terms of public health and its cost estimate (economic impact). CURRENT STATE OF KNOWLEDGE The problem of diabetic ketoacidosis has been reviewed recently (38). A brief summary of the pathophysiology and clinical picture can be constructed as follows. Diabetic coma is initiated by a deficiency of insulin which is more profound than that found in the uncomplicated 123 124 Diabetes Mellitus diabetic state. Very recent studies (49), utilizing a connecting peptide immunoassay to assess insulin secretion, have shown essentially no insulin secretion during the period of ketoacidosis, while insulin release (at least in some patients) subsequently may return to measurable levels. Concomitant with insulin deficiency, there appear to be changes in other hormones with increased plasma concentrations of cortisol, growth hormone, and glucagon being regularly observed (28, 43). Associated with the hormonal changes there is an activation of gluconeogenesis, which, when coupled with a diminished peripheral utilization of glucose, results in marked hyper- glycemia with its consequence of osmotic diuresis, volume depletion of body fluids, osmolar concentration, and ultimately shock and renal shutdown. At the same time ketone body formation begins to increase and progresses to the severe metabolic acidosis which characterizes diabetic coma. Essentially all of the clinical aspects of the ketoacidotic picture are directly trace- able to these two processes and their interrelationships. The mechanisms whereby they are regulated thus assume major importance. a. Gluconeogenesis: The control of gluconeogenesis in starvation and diabetes has been the subject of enormous study but remains a matter of controversy. The problem, as can be seen from Exton's review (22) of the subject, is that control has been postulated for many different sites and evidence can be adduced to support each of these. As emphasized by Srere (52) such data, particularly that obtained in vitro, may have little relation to the situation operative in the intact organism. Doubtless all major pathways of intermediary metabolism are under multiple interlocking controls to assure normal function. It follows that in disease states abnormalities likely exist at multitudinous points. Nevertheless, in the broad sense, it can be stated that gluconeogenesis, under the influence of the hormonal changes mentioned above, is activated both by the provision of increased quantities of substrate and by acceleration of certain key enzymatic reactions of the gluconeogenic pathway in the liver and perhaps the kidney. Thus the following general formulation can be put forth: Hormone changes —— Mobilization of substrate Activation of enzymes — Increased product formation In the case of gluconeogenesis the substrates mobilized are amino acids, glycerol, lactate, and pyruvate (17,22,23). The uptake of the amino acids in the liver may also be hormonally controlled, though evidence for such a process with the other substrates is lacking. From the standpoint of enzyme activation, attention has focused on four sites in the pathway from pyruvate to glucose; pyruvate carboxylase, phosphoenolpyruvate carboxykinase, fructose-1, 6-diphosphatase, and glucose-6-phosphatase. Activities of these enzymes appear to increase in concert when gluconeogenesis is activated. It is also thought possible (or likely) that the counterpart enzymes of the glycolytic pathway (e.g., phosphofructokinase, pyruvate kinase, pyruvate dehydrogenase) are inhibited under these circumstances. These changes can apparently occur very rapidly (4-5 minutes) under certain circumstances (53). No firm insight is available as to the mechanism of activation of gluconeogenic enzymes, though both allosteric regulation and increased enzyme synthesis are presumed to play a role. Whatever the mechanism, the end Diabetic Coma 125 result is a marked overproduction of glucose by the liver: in diabetic coma hepatic glucose release may reach 1000 G a day, some 3 times higher than the maximum rate attained in starvation 12, 17). b. Ketogenesis: The general formulation given for enhanced gluconeogenesis in diabetic coma is also applicable to the increase in ketone body formation. Here again the process is thought to be the result of hormonal changes which cause mobilization of substrate to the liver and which alter the metabolism of that organ by changing enzyme activity. In this case the substrate mobilized is free fatty acid from peripheral fat stores while the products are acetoacetic and B-hydroxybutyric acids. The enzymic sequence primarily involved is the B-oxidation pathway for fatty acids (39). Under normal circumstances free fatty acids taken up by the liver are utilized primarily for triglyceride synthesis, while relatively little is oxidized to co, or ketone bodies. During starvation or diabetes, on the other hand, a very significant fraction of the fatty acids taken up enter the oxidative pathway. In contrast to the situation with gluconeogenesis, control of ketogenesis may be primarily vested in a single enzymatic reaction, the long chain fatty acylcarnitine transferase reaction. Compelling evidence for this conclusion is the demonstration that octanoic acid, which does not require a carnitine mechanism for transfer into the mitochondrion, is oxidized at similar rates by livers from normal, starved, and diabetic animals. Moreover, while reversal of ketosis by insulin in vivo cannot be overcome by infusing long chain fatty acids (8) the infusion of octanoate immediately restores ketone production to the previous rate (36). Such findings have led to the concept that the capacity for B-oxidation in the mitochondrion is fixed and high and that the rate-limiting step for the sequence is the transport of the long chain fatty acid through the mitochondrial membrane. Once this is accomplished, oxidation to acetyl CoA occurs rapid- ly with the subsequent formation of ketone bodies. When flux through the pathway is rapid, the utilization of acetyl CoA in the Krebs cycle and for fatty acid synthesis is minimal and the bulk of the acetyl CoA is converted to acetoacetate and B-hydroxybutyrate. Accordingly, significant importance has been attached to the possible role of the long chain acylcarnitine transferase reaction, though to the present little information is available regarding its control (44). It should be noted that the removal (utilization) system for ketone bodies in peripheral tissues appears to be saturable and that the final level of plasma ketones is the consequence of both increased hepatic production and impaired utilization (4,5,35). e. Clinical picture and treatment: Since the clinical details of diabetic ketoacidosis are well known, it will not be described in detail. The composite picture is one of hyperglycemia, intracellular and extracellular fluid volume deficit with osmolar concentration, total body potassium and phosphate depletion, and severe metabolic acidosis. Therapy, which is generally successful, consists of the administration of large amounts of insulin and saline infusions with careful attention to serum potassium concentrqtions (38). INFORMATION NEEDED TO BE ACQUIRED THROUGH RESEARCH No attempt will be made to outline specific and detailed experiments in this section or to exhaustively list unsolved problems. However, certain major questions needing to be addressed will be indicated as representative problems under four major categories: 126 Diabetes Mellitus The physiology of insulin and its mechanism of action, The control of gluconeogenesis, 0 op The control of ketogenesis, d. Clinical problems. a. The physiology of insulin and its mechanism of action: Few hormones have received the extensive study accorded insulin. While a great deal has been learned regarding its effects in man and various animal species, a number of fundamental questions remain that are immediately related to the problem of diabetic coma. (1) How does insulin work in the liver? This is the most basic of all questions. There is little doubt that insulin has a direct effect on hepatic metabolism in vivo which is powerful and rapid (26 41) yet effects in vitro are extremely difficult or impossible to demonstrate. It is clear that insulin binds to hepatic plasma membranes (19) as it does to other tissues, and it has been possible to show effects of insulin on glycogen synthesis and gluconeogenesis (30) in the perfused rat liver. The changes are of small degree, however, and often restricted to rigidly defined conditions. In regard to ketogenesis the record is even more blank. In many experiments under widely varying conditions, Dr. McGarry and I have never been able to show a depressant effect of insulin on ketone body production in vitro (37). While experiments should continue to be directed at biochemical interrelationships of insulin such as with the cyclic AMP system and membrane enzymes like the oubain inhibitable ATPase (11), it seems clear that another fruitful area for investigation would be hepatic structural changes produced by insulin (47). Finally, consideration should be given to the possibility that the failure to demonstrate insulin effects on the liver in vitro might be related to a marked activation of insulin degrading systems (20). Major effort should be expended on these problems. (2) What is the nature of insulin resistance in diabetic ketoacidosis? It is a well-known clinical observation that small amounts of insulin normally suppress hepatic ketone body production even under circumstances where hyperglycemia is not controlled. On the other hand, in diabetic acidosis very large quantities of insulin are ineffective in rapidly reversing ketogenesis. It is not known whether this resistance is due to plasma factors (24) or to changes in the liver itself. (3) Why does insulin release stop in ketoacidosis? As mentioned above, Rubenstein and coworkers have shown that insulin release becomes immeasurable in ketoacidosis. The mechanism is totally unknown. Since a variety of stresses are known to precipitate ketoacidosis, the presumption would be that the cut off is a final common pathway of stress, possibly hormonal in nature. A number of mechanisms are already recognized for the normal control of insulin secretion (46) but to the present careful evaluation of control phenomena in ketoacidosis has not been carried out. The catecholamines would appear to be likely candidates here, but attention should also be given to interaction with other hormones (such as gastrin) which seem to function primarily as facilitators of the normal beta cell response to glucose (48). Similarly, it would be extremely important to know if the synthetic pathway for insulin (54) is altered during the acute episode of ketoacidosis (or immediately preceding it). At the present there is no possibility of carrying out these studies in humans, requiring that an Diabetic Coma 127 animal model be used. The latter is complicated by the fact that ketoacidosis does not occur spontaneously and would have to be induced. (4) Is the insulin:glucagon ratio a unique function? Recently a number of investiga- tors, following the suggestion of Unger (55), have indicated that the insulin:glucagon ratio in plasma may play a distinct role in uncontrolled diabetes; i.e., rather than considering the abnormalities of diabetic coma as the integrated result of multiple hormonal changes, the hypothesis has been put forward that a unique role is played by the combination of insulin deficiency and glucagon excess. This issue, while difficult, can be approached in the experi- mental animal utilizing antiglucagon antibodies during the induction of diabetic ketoacidosis or through the development of specific inhibitors of glucagon release. The latter compounds could have clinical usefulness. Preliminary studies in this area clearly need to be expanded. b. The control of gluconeogenesis: Since gluconeogenesis, as mentioned above, contributes in a major way to the hyperglycemia of diabetic coma and other forms of uncontrolled diabetes, broader understanding of its control is needed. (1) Does glucagon stimulate gluconeogenesis in man? There seems to be little question that glucagon stimulates gluconeogenesis in the rat (9, 10). On the other hand, Madison (42) has been unable to confirm a gluconeogenic effect in the dog. Definitive information on this point is sorely needed. If glucagon enhances gluconeogenesis, the mechanism must be further explored. From studies in rat liver, it has been suggested that glucagon only works in the presence of adrenal corticosteroids (21) but the precise site of action remains uncertain (10). (2) How is gluconeogenesis controlled? As mentioned above, a great deal of work remains to be done regarding the control of gluconeogenesis. For example, the once quite popular theory of regulation by the redox state of the cell has come to be doubted (29) and current work emphasizes the role of substrate phosphorylation, particularly GTP formation (16). The critical issue is to discern the operating features in vivo, since so many mechanisms appear to be operative in vitro. At the present time no good techniques exist for correlating in vitro and in vivo findings and new methods will obviously have to be developed. (3) What is the relationship of gluconeogenesis to ketogenesis? Both of these processes are accelerated in diabetic coma, and the general rule is that when flux is acceler- ated over one pathway, it is also increased in the other. Indeed, Flatt (25) has suggested that ketogenesis is limited by the rate of gluconeogenesis. On the other hand, there are a number of circumstances in which the two can be separated. Preeminent here would be the in- fusion of lactate to the isolated perfused liver from ketotic animals; this substrate stimulates gluconeogenesis but inhibits ketogenesis (36). Of even greater interest is the observation (McGarry and Foster, unpublished) that alcohol, which is potently antiketogenic in the presence of active gluconeogenesis, is totally ineffective when gluconeogenesis is blocked. Dissection of these interrelationships may provide clues to the fundamental mechanism through which carbohydrate and lipid metabolism are linked. (This relationship is seen in classic fashion when the accelerated fat metabolism and ketosis of starvation are reversed by carbohydrate feeding). e. The control of ketogenesis: Because of the close interrelationship between the mechanism of insulin action, the control of gluconeogenesis, and the regulation of ketogenesis, several problems relating to ketogenesis have already been mentioned. Other representative 128 Diabetes Mellitus areas needing exploration are identified by the following questions: (1) How is fatty acid oxidation controlled? As has already been described, ketone body formation is essentially regulated by the rate of fatty acid oxidation, which in turn appears to be controlled directly or indirectly by the presence of glucose and insulin. Since glucose itself is not antiketogenic in vitro, while products or precursors such as lactate and dihydroxyacetone are (37, 58) an attractive possibility is that some glycolytic-gluconeogenic intermediate acts allosterically to activate the beta oxidation sequence. For reasons outlined earlier in the discussion, the prime candidate for the site of such control is the long chain acylcarnitine transferase reaction. While considerable new information regarding this enzyme is available (15,33), basic control mechanisms remain to be elucidated. (2) What controls hepatic lipase activity? As emphasized earlier, accelerated ketogenesis can only be sustained if the supply of substrate, fatty acid, is maintained high. It is well known that the diabetic liver has an increased triglyceride content and that the isolated perfused liver from diabetic animals continues to make ketone bodies at a high rate in the absence of exogenous substrate (37,56). Since the hormone sensitive lipase of adipose tissue is rapidly inhibited by insulin (32), while lipase activity in the liver is not (at least in the perfused liver), it follows that study of the hepatic enzyme is imperative. Indeed, since there are several hepatic lipases (3), the enzyme coupled with ketosis will have to be identified. (3) What limits ketogenesis in hyperosmolar coma and generalized or partial lipodystrophy? The absence of ketosis in hyperosmolar coma remains an enigma (27). Similarly, the absence of ketosis in the lipodystrophies needs to be explained (51). The problem is dif- ficult, but in the former state portal vein insulins need to be measured. Since peripheral blood insulin values are equivalent in diabetic ketoacidosis and hyperosmolar coma, both types of patients need to have ketone response measured after ingestion of medium chain triglycerides as a source of octanoic acid. This would indicate whether or not the ketogenic machinery itself is intact. (4) 1Is there a role for inhibitors of fatty acid oxidation in human ketoacidosis? McGarry and Foster (40) have reported the rapid reversal of experimental diabetic ketoacidosis in the rat with (+)-decanoylcarnitine. The importance of this observation lies in the fact that recovery times with (+)-decanoylcarnitine may be one-half or less those achievable with insulin alone. More recently we have synthesized a homologue of (+)-decanoylcarnitine which is equally potent in reversing ketoacidosis but which has remarkably diminished surface active properties that should make it much safer for in vivo use. The possibility of utilizing pharmacologic agents of this sort in human diabetic coma has much potential. Studies in this area clearly need to be expanded. d. Clinical problems: The following are examples of clinical problems that need to be addressed. (1) What is the cause of central nervous system dysfunction in diabetic coma? It is a remarkable fact that despite many years of study, the nature of the central nervous system depression in ketoacidosis remains unknown. The interrelationships between CSF pH (45), cerebral oxygen uptake (31) and direct ketone toxicity (50) all need to be carefully reexamined. (2) What is the mechanism of cerebral edema during treatment of diabetic acidosis? Diabetic Coma 129 The now well-recognized syndrome of cerebral edema associated with reversal of human ketoacidosis (60) is thought to be associated with osmotic disequilibrium between plasma and brain during the rapid lowering of the blood sugar (18). However, precise mechanisms have not been worked out and the possibility of direct membrane damage is intriguing (2). The area is ripe for further in- vestigation. (3) Should phosphate solutions be used routinely in diabetic coma? In view of the depletion of red cell 2,3-diphosphoglycerate in diabetic ketoacidosis (7), it has been suggested (38, 1) that phosphate salts might be useful in treatment. While there are sound physiologic arguments to support this suggestion, controlled clinical studies are needed to test the idea. APPLICATION TO PREVENTIVE MEDICINE It is obvious that the complications of uncontrolled diabetes, such as diabetic coma, are enormously costly from the standpoint of morbidity, mortality, and economics. Studies of the type outlined should, over the short range, provide new and improved forms of treatment for diabetic ketoacidosis. I believe this will happen within the next 3 years. From a long-range standpoint, basic understanding of the pathophysiology may lead to ways of completely preventing the onset of diabetic coma. SUMMARY Diabetic coma is a serious clinical problem about which much remains to be learned. It is not necessary to search for new areas of research, since so many basic questions are recognized and are obviously in need of answers here and now. This chapter is intended to indicate this fact. REFERENCES 1. Alberti, KGMM, JH Darley, PM Emerson, and TDR Hockaday 1972. 2,3-Diphosphoglycerate and tissue oxygenation in uncontrolled diabetes mellitus. Lancet 2:391. 2. Arieff, AI, and CR Kleeman 1973. Studies on mechanisms of cerebral edema in diabetic comas. Effects of hyperglycemia and rapid lowering of plasma glucose in normal rabbits. J Clin Invest 52:571. 3. Assmann, G, RM Krauss, DS Frederickson, and RI Levy 1973. Characterization, subcellular localization, and partial purification of a heparin-released triglyceride lipase from rat liver. J Biol Chem 248:1992. 4. Balasse, EO, and RJ Havel 1970. Turnover rate and oxidation of ketone bodies in normal and diabetic dogs. Diabetologia 6:36. 5. Bates, MW, HA Krebs, and DH Williamson 1968. Turnover rates of ketone bodies in normal, starved, and alloxan diabetic rats. Biochem J 110:655. 6. Beigelman, PM 1971. Severe diabetic ketoacidosis (diabetic ''coma'). 482 episodes in 257 patients, experience of three years. Diabetes 20:490. 7. Bellingham, AJ, JC Detter, and C Lenfant 1970. The role of hemoglobin oxygen affinity and red cell 2,3-DPG in the management of diabetic ketoacidosis. Trans Ass Amer Phys 83:113. 8. Bieberdorf, FA, SS Chernick, and RO Scow 1970. Effect of insulin and acute diabetes on plasma FFA and ketone bodies in the fasting rat. J Clin Invest 49:1685. 130 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24, 25. 26. 27. 28. 30. 31. Diabetes Mellitus Blair, JB, DE Cook, and HA Lardy 1973a. Influence of glucagon on the metabolism of xylitol and dihydroxyacetone in the isolated perfused rat liver. J Biol Chem 248:3601. Blair, JB, DE Cook, and HA Lardy 1973b. Interaction of propionate and lactate in the perfused rat liver. J Biol Chem 248:3608. Blatt, LM, PH McVerry, and K-H Kim 1972. Regulation of hepatic glycogen synthetase of Rana catesbeiana. Inhibition of the action of insulin by oubain. J Biol Chem 247:6551. Bondy, PK, WL Bloom, VS Whitner, and BW Farrar 1949. Studies of the role of the liver in human carbohydrate metabolism by the venous catheter technique. II. Patients with diabetic ketoacidosis before and after the administration of insulin. J Clin Invest 28:1126. Boucot, KR, DA Cooper, ES Dillon, P Meier, and R Richardson 1952. Tuberculosis Among the Diabetic. The Philadelphia Survey. American Review of Tuberculosis, Volume 65(suppl) 1-50. Bressler, R 1966. The biochemistry of ketosis. Ann N Y Acad Sci 104:735. Brosnan, JT, B Kopec, and IB Fritz 1973. The localization of carnitine palmitoyltransferase on the inner membrane of bovine liver mitochondria. J Biol Chem 248:4075. Bryla, J, CM Smith, and JR Williamson 1973. Control of phosphoenolpyruvate synthesis by substrate level phosphorylation in guinea pig liver mitochondria. J Biol Chem 248:4003. Cahill, GF, Jr 1971. Physiology of insulin in man. Diabetes 20:785. Clements, RS, Jr, LD Prockop, and AI Winegrad 1968. Acute cerebral oedema during treatment of hyperglycemia. Lancet 2:384. Cuatrecasas, P 1973. Interaction of concanavalin A and wheat germ agglutinin with the insulin receptor of fat cells and liver. J Biol Chem 248:3528. Duckworth, WC, MA Heinemann, and AE Kitabchi 1972. Purification of insulin-specific protease by affinity chromatography. Proc Nat'l Acad Sci 69:3698. Exton, JH, N Friedmann, EH Wong, JP Brineaux, JD Corbin, and CR Park 1972. Interaction of glucocorticoids with glucagon and epinephrine in the control of gluconeogenesis and glycogenolysis in liver and lipolysis in adipose tissue. J Biol Chem 247:3579. Exton, JH 1972. Gluconeogenesis. Metabolism 21:945. Felig, P 1973. The glucose-alanine cycle. Metabolism 22:179. Field, JB 1962. Insulin resistance in diabetes. Ann Rev Med 13:249. Flatt, JP 1972. On the maximal possible rate of ketogenesis. Diabetes 21:50. Foster, DW 1967. Studies in the ketosis of fasting. J Clin Invest 46:1283. Foster, DW 1973. Insulin deficiency and hyperosmolar coma. Adv Int Med (in press). Gerich, JE, MM Martin, and L Recant 1971. Clinical and metabolic characteristics of hyperosmolar nonketotic coma. Diabetes 20:228. Hawkins, RA, CRS Houghton, and DW Williamson 1973. Hepatic redox state and gluconeogenesis from lactate in vivo in the rat. Biochem J 132:19-25. Jefferson, LS, JH Exton, RW Butcher, EW Sutherland, and CR Park 1968. Role of adenosine- 3'5'-monophosphate in the effects of insulin and anti-insulin serum on liver metabolism. J Biol Chem 243:1031. Kety, SS, BD Polis, CS Nadler, and CF Schmidt 1948. The blood flow and oxygen consumption of the human brain in diabetic acidosis and coma. J Clin Invest 27:500. 32. 33. 34. 35. 36. 37. 33. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. Diabetic Coma 131 Khoo, JC, D Steinberg, B Thompson, and SE Mayer 1973. Hormonal regulation of adipocyte enzymes. The effects of ephinephrine and insulin in the control of lipase, phosphorylase kinase, phosphorylase, and glycogen synthesis. J Biol Chem 248:3823. Kopec, B, and IB Fritz 1973. "Comparison of properties of carnitine palmitoyltransferase I with those of carnitine palmitoyltransferase II, and preparation of antibodies to carnitine palmitoyltransferases. J Biol Chem 248:4069. Krebs, HA 1966. The regulation of the release of ketone bodies by the liver. Advances Enz Regulat 4:339. McGarry, JD, MJ Guest, and DW Foster 1970. Ketone body metabolism in the ketosis of starvation and alloxan diabetes. J Biol Chem 245:4382. McGarry, JD, and DW Foster 197la. The regulation of ketogenesis from octanoic acid. The role of the tricarboxylic acid cycle and fatty acid synthesis. J Biol Chem 248:1149. McGarry, JD, and DW Foster 1971b. The regulation of ketogenesis from oleic acid and the influence of antiketogenic agents. J Biol Chem 246:6247. McGarry, JD, and DW Foster 1972. Regulation of ketogenesis and clinical aspects of the ketotic state. Metabolism 21:471. McGarry, JD, JM Meier, and DW Foster 1973. The effects of starvation and refeeding on carbohydrate and lipid metabolism in vivo and in the perfused rat liver. The relationship between fatty acid oxidation and esterification in the regulation of ketogenesis. J Biol Chem 243:270. McGarry, JD, and DW Foster 1973. Acute reversal of experimental diabetic ketoacidosis in the rat with (+)-decanoylcarnitine. J Clin Invest 52:877. Madison, LL 1969. Role of insulin in the hepatic handling of glucose. Arch Int Med 123:284. Madison, LL 1973. Personal communication. Miller, WA, GR Faloona, E Aguilar-Parada, and RH Unger 1970. Abnormal alpha-cell function in diabetes. Response to carbohydrate and protein ingestion. N Eng J Med 283:109. Norum, RN 1965. Activation of palmityl-CoA:carnitine palmityltransferase in livers from fasted, fat-fed, or diabetic rats. Biochem Biophys Acta 98:652. Ohman, JL, Jr, EB Marliss, TT Aoki, CS Munichoodoppa, VK Khanna, and GP Kozak 1971. The cerebrospinal fluid in diabetic ketoacidosis. N Eng J Med 284:283. Porte, D, Jr, and JD Bagdade 1970. Human insulin secretion: an integrated approach. Ann Rev Med 21:219. Reaven, EP, DT Peterson, and GM Reaven 1973. The effect of experimental diabetes mellitus and insulin replacement on hepatic ultrastructure and protein synthesis. J Clin Invest 52:248. Rehfeld, JF, and F Stadil 1973. The effect of gastrin on basal and glucose-stimulated insulin secretion in man. J Clin Invest 52:1415. Rubenstein, AH 1973. Personal communication. Schneider, R, and H Droller 1938. Relative importance of ketosis and acidosis in production of diabetic coma. Quart J Exper Physiol 28:323. Seip, M 1971. Generalized lipodystrophy. Ergeb Inn Med Kind 31:59. Srere, PA 1969. Some complexities of metabolic regulation. Biochem Med 3:61. 132 53. 54. 55. 56. 57. 58. 59. 60. Diabetes Mellitus Stifel, FB, OD Taunton, HL Greene, and RH Herman 1973. Rapid changes of rat hepatic and extrahepatic glycolytic enzymes and fructose-1,6,-diphosphatase following glucagon injection in vivo. Fed Proc 32:897 abs. Tager, HS, SO Emdin, JL Clark, and DF Steiner 1973. Studies on the conversion of proinsulin to insulin. II. Evidence for a chymotrypsin-like cleavage in the connecting peptide region of insulin precursors in the rat. J Biol Chem 248:3476. Unger, RH 1972. Circulating pancreatic and extra-pancreatic glucagon-like materials 15] is not diagnostic of lactic acidosis but may also be seen with renal failure, sulleyinte itorieation, methanol poisoning, ethylene glycol ingestion, and diabetic ketoacidosis. If lactic acidosis is superimposed on or associated with an acidosis characterized by an increased "anion gap,' the diagnosis may be difficult. This is particularly true when there is coexistent diabetes mellitus and ketoacidosis. At the present time, there is no way to make the diagnosis of lactic acidosis in the presence of ketoacidosis without a blood lactate measurement. On the other hand, the presence of only trace amounts of plasma ketones does not necessarily indicate lactic acidosis since significant ketoacidosis can exist in the presence of weakly positive plasma acetest reactions. The failure of a patient with ketoacidosis to improve despite adequate therapy or the persistence of acidosis which is resistant to alkali should suggest coexistent or superimposed lactic acidosis. The onset of lactic acidosis is traditionally described as precipitous, particulary when it occurs in the hospital as a consequence of severe underlying disease and altered tissue perfusion. Whether the onset of idiopathic lactic acidosis or lactic acidosis in patients receiving phenformin is acute, is not known. Our experience, and that of others with diabetic patients who are found in the emergency room to have idiopathic lactic acidosis, is that a history of prodromal symptoms developing over several days can often be obtained (62). This suggests that diabetics, particularly those receiving phenformin, should be closely observed for the development of the nonspecific symptoms and if present and unexplained, the possibility that subtle lactic acidosis is present and/or developing must be considered. DIABETES MELLITUS AND LACTIC ACIDOSIS The occasional occurrence of nonketotic acidosis and coma in patients with diabetes mellitus has been known for many years (44). Appel and Cooper (5) reviewed the subject and attributed the acidosis to impaired renal excretion of ketones. Lexow (36) described three diabetic patients with coma and acidosis without ketosis and suggested that the acidosis was due to lactic acid. However, it was not until Daughaday (14) described an additional three diabetic patients with 146 Diabetes Mellitus nonketotic acidosis, that excessive accumulation of lactic acid could be established as the cause for the acidosis. Because of the frequent occurrence of diabetes in patients with idiopathic lactic acidosis, a causal relationship between diabetes mellitus and lactic acidosis has been suggested. There are no obvious metabolic defects in experimental diabetes mellitus in animals (22) nor spontaneous diabetes mellitus in humans which predispose to the development of lactic acidosis. Blood lactate concentrations in 83 acutely ill patients with diabetes studied by Traquada (62) were normal or slightly elevated with a mean value of 1.22 mM. Although Shreeve (50) observed de- creased lactate-!%C clearance in diabetes mellitus lactate disposal, as reflected by the rate of removal of an exogenous oral lactate load, it is not prolonged (4). The administration of glucose to normal subjects and insulin to patients with diabetes results in increased glycolysis but only a slight increase in blood lactate concentrations (27, 57). The incidence of lactic acidosis in patients with ketoacidosis seems relatively small. In a series of 12 patients studied by Strangaard (55) blood lactate and pyruvate concentrations were within normal limits and lactate increased only slightly in some patients with treatment. Further- more, these investigators observed an inverse linear relationship between total blood ketone and bicarbonate concentrations indicating the absence of lactate and other unmeasured anions. In another series of 23 patients with diabetic ketoacidosis studied by Watkins (67) the contribution of lactate to the anion gap and the acidosis was small and an excellent correlation was also observed between the concentration of B-hydroxybutyrate and pH, again emphasizing the unimportance of lactate in this study. However, in seven of these patients, the blood lactate concentrations were in excess of 3.0 mM and in three patients were 4.2, 7.3, and 19.4 mM, respectively. Al- though blood lactate concentrations were, with the three exceptions mentioned, either within normal limits or only minimally elevated blood lactate concentrations increased during therapy of the ketoacidosis in 13 of 17 patients in whom it was measured, two of whom died. One of the latter had a severe predisposing disorder that ordinarily would be associated with lactic acidosis and invocation of diabetes mellitus is unnecessary. In two patients in this series, the acidosis appeared to be due to lactic acid. Alberti and Hockaday (1) have reported lactate and pyruvate measurements in a series of 50 patients with diabetic ketoacidosis. The mean lactate and pyruvate concentrations were 2.49 and 0.14 mM, respectively, and the L/P ratio was 19. Based upon the changes that occurred in blood lactate with insulin treatment, these patients were subdivided into two groups. Approximately one-third of the patients demonstrated a fall in blood lactate, the magnitude of which inversely correlated with the initial lactate concentration (i.e., the higher the concentration, the greater the fall). This group was characterized by higher initial lactate concentrations and greater degrees of hyperglycemia, hyperketonemia, and acidosis. They also required more insulin than did the other group, suggesting that the diabetic ketoacidosis was more severe. In the remaining patients, the blood lactate concentrations transiently in- creased during therapy. In none, however, was severe or progressive lactic acidosis detected. Although there is a gradation of lactate concentrations in diabetic ketoacidosis, these authors have observed only four patients in a series of 55 in whom the lactate was greater than 7.0 mM (26). It would thus appear that significant lactic acidosis is not a common problem in patients with ketoacidosis nor does insulin therapy, despite its ability to increase blood lactate concen- trations in normal and diabetic subjects result in lactic acidosis. Lactic Acidosis 147 Most diabetics with lactic acidosis are in shock or have severe underlying diseases or dis- orders known to be associated with the syndrome (62). Nonetheless, there are patients with lactic acidosis, in the absence of known predisposing factors, in whom diabetes mellitus or some disturbance of carbohydrate metabolism is present (38, 62). It is not clear whether such ab- normalities in carbohydrate metabolism are etiologically related to the acidosis or just co- incidental. The prognosis of patients with idiopathic lactic acidosis, with and without diabetes mellitus, is poor. In Oliva's review article on lactic acidosis (38) only 5 of 15 patients de- scribed in the literature with both lactic acidosis and diabetes mellitus had survived while the mortality rate was 100 percent in patients with lactic acidosis in the absence of diabetes. In a series of 58 patients, both with and without diabetes mellitus, reported by Tranquada (60), the mortality rate was 90 percent. The studies of blood lactate concentrations in patients with ketoacidosis do not indicate the presence of metabolic lesions that predispose to the development of lactic acidosis. However, because these patients are younger and therefore less likely to have vascular disease, the possibility remains that there may be an anatomic lesion which, when associated with diabetes predisposes to the development of this syndrome. The available information concerning lactic acidosis and coexistent diabetes mellitus reveals that the patients are generally older and demonstrate only mild to modest hyperglycemia despite severe stress. These individuals could perhaps be best described as having maturity onset diabetes mellitus and may be more at risk because of age and accompanying vascular disease, which is more extensive than in nondiabetics, rather than because of any biochemical predisposition. PHENFORMIN AND LACTIC ACIDOSIS The role of phenformin in lactic acidosis is highly controversial and the high incidence of diabetes mellitus in patients with lactic acidosis who are not receiving phenformin confounds the issue and makes the definition of this relationship difficult. Shortly after its introduction, attention was drawn to the occurrence of severe and fatal metabolic acidosis in a series of dia- betics who were receiving phenformin (65). Subsequent publications (8, 10, 12, 15, 59) have in- ferred a direct relationship between phenformin and lactic acidosis. In 1973 there were approxi- mately 200 cases (documented in the scientific literature or reported directly to the manufacturers) of lactic acidosis occurring during phenformin treatment (63). Analysis of these reports reveal that 70 percent of the involved individuals were females, 55 percent were between the ages of 60 and 80 years, and 90 percent were older than 41 years. Virtually all patients were receiving between 100-150 mg per day. These patients were found to differ in a number of ways when compared with a randomly selected series of 181 patients with lactic acidosis who were not receiving phenformin. The mean age of the patients who were receiving phenformin was 61 years in contrast to 51 years for the nonphenformin group. In 58 percent of the phenformin series, the lactate concentration was greater than 6.5 mM and the pH was less than 7.30-7.33. Known causes of lactic acidosis were present in 56 percent of the cases and in an additional 17 percent, conditions were present which are known to contribute to the development of lactic acidosis. In only 30 percent of the group receiving phenformin was lactic acidosis present without known or contributing causes. Recovery occurred in 50 percent of the phenformin-treated group, but in only 22 percent of the nonphenformin group. In a series of 21 patients with lactic acidosis who were receiving phenformin, described by Bengtsson (8) 67 percent survived. However, the survival rate in the nine patients described by Cleaver and Carretta (12) was only 22 percent. 148 Diabetes Mellitus The incidence of lactic acidosis in phenformin-treated diabetes is unknown but considering that 400,000 patients were estimated as receiving phenformin in 1972, the occurrence must be rela- tively rare. In opposition Bengtsson and co-workers observed 21 patients in a large diabetic clinic over a period of 39 months in whom lactic acidosis developed during oral antidiabetic treatment in which phenformin was being used alone or in combination with sulfonylurea agents or insulin (8). Cleaver and Carretta (12) have observed nine cases of lactic acidosis in diabetic patients receiving phenformin and although they have claimed that there is a more common problem than previously appreciated, the time interval over which their cases were accumulated was not stated. The intense concern over the potential relationship of phenformin to lactic acidosis is a direct result of observations in animals and humans that it can increase lactate production and/ or elevate the blood lactate concentration. In initial in vitro studies, phenformin was shown to inhibit oxidative phosphorylation, increase anaerobic metabolism and the production of lactic acid in skeletal muscle (51). Subsequently, it has been demonstrated to inhibit gluconeogenesis from lactate in rat liver slices (41), renal cortex (41) and in isolated perfused rat and guinea pig livers (3, 58), indicating that the hyperlactatemia in intact animals may be theoretically due to both increased production and decreased utilization. Although the effects of phenformin on cellular metabolism Zn vitro have often been demonstrated at concentrations considerably in excess of those seen therapeutically, the observations cannot be dismissed because certain tissues (skeletal muscle, liver and the gastrointestinal tract) concentrate phenformin (25) and tissue levels of the drug may be greatly in excess of that expected from plasma concentrations. More recently, inhibition of gluconeogenesis by rat liver has been achieved with concentrations of phenformin that are within the therapeutic range (58). In normal human volunteers, therapeutic doses of phenformin have produced mild elevations of blood lactate and have accelerated glucose turnover and its conversion to lactate while also increasing the turnover and incorporation of lactate into glucose (31, 32). Similar changes have also been observed in patients with diabetes mellitus (47). Increased glucose conversion to lactate may simply reflect enhanced glycolysis and does not imply an increase in anaerobic metabolism. Increased concentrations of blood pyruvate in phenformin-treated subjects (13) and normal or slightly increased glucose oxidation (31, 47) supports such a proposal. Dembo and associates (16) have suggested that the changes in pyruvate metabolism that accom- pany relative insulin deficiency may, in the presence of phenformin, lead to lactic acidosis. In this scheme insulin deficiency and fasting result in the overproduction and underutilization of pyruvate. The former from enhanced protein breakdown and provision of three carbon precursors and the latter from preferential utilization of fat and inhibition of pyruvate oxidation. The shift to a more reduced cellular redox state as a consequence of increased fatty acid oxidation may be further accentuated by the accumulation of excessive amounts of phenformin which occurs with renal impairment and its further effect on cellular redox. Recently, Searle (48) has studied the metabolism of lactate-l"C in a patient who developed lactic acidosis with the adminis- tration of phenformin and demonstrated that lactate oxidation did not increase sufficiently to keep pace with the increased rate of lactic acid formation that occurred. He suggests that the imbalance in lactate metabolism is responsible for the lactate acidosis that accompanies phenformin therapy. Lactic Acidosis 149 Because of the high incidence of diabetes mellitus in patients with idiopathic lactic acidosis, it is particularly difficult to determine whether phenformin exerts an independent effect, syner- gizes in some way with the abnormality in carbohydrate metabolism present in diabetes, or is co- incidental and unimportant as a cause of lactic acidosis. The possibility must be seriously considered; however, that under certain circumstances phenformin predisposes to the development of lactic acidosis. It may be speculated that in the presence of marginal tissue oxygenation, phenformin may increase or potentiate lactic acid production beyond that which would have occurred, thus converting a situation in which compensation could be maintained or spontaneously reversed into progressive acidosis. Studies in rats on this point are conflicting. Lacher (35) demon- strated impairment in the disposal of both endogenous and exogenous lactate loads in phenformin- treated rats, while Ruggles (46) was unable to show potentiation of hyperlactatemia resulting from hypoxemia. In humans, phenformin does not potentiate exercise-induced hyperlactatemia (23, 49). Other predisposing factors, particularly ethanol, should not be minimized since ethanol and phen- formin act synergistically on blood lactate concentrations (34). Phenformin is eliminated from the body by the kidneys (6) and in the presence of renal disease impaired excretion and elevation of blood levels of phenformin may be expected. This is important since there is a rough relationship between the dose of phenformin used in maturity onset diabetic patients and the increase observed in the blood lactate concentration. Hyperlactatemia does not occur with phenformin doses of less than 75 to 100 mg daily, while greater increments in lactate are observed with doses of 175 to 250 mg daily (13, 19). In view of the relationship between the dose of phenformin and the blood lactate concentration, it is reasonable that impaired excretion of phenformin results in drug concentrations that produce or predispose to lactic acidosis. The development of lactic acidosis with suicidal overdoses of phenformin would also suggest that blood concentations may be an important factor in the genesis of this syndrome (15, 43). The vast ma- jority of patients who develop lactic acidosis while receiving phenformin have evidence of impaired renal function (62, 63). There may be certain individuals with mild or modest degrees of impair- ment in renal function in the absence of azotemia, in whom alterations in hydration alter renal perfusion and result in the retention of phenformin. Consequently, it seems justifiable to recom- mend that phenformin not be used in patients who have evidence of renal impairment and that creatinine clearance measurements be considered in all patients with obvious renal disease pre- liminary to the initiation of therapy with this drug. TREATMENT There is very little that can be said concerning the treatment of lactic acidosis. Recovery is usually determined by the response of the basic underlying disorder to specific therapy. Bi- carbonate treatment is not successful, even when correction of pH can be obtained, unless the basic defects can be identified and corrected. Unfortunately, even when the underlying pathology can be defined, the disorder may not be reversible and the prognosis for survival is poor. Since the pathogenesis of idiopathic lactic acidosis is not known, no specifically directed therapy is possible. Recovery of patients with idiopathic lactic acidosis is uncommon, unpredictable and in most instances when it occurs, unexplained. The mortality of both nondiabetic and diabetic patients with idiopathic lactic acidosis is high (28, 30, 62). Patients who develop lactic acidosis while receiving phenformin may have a somewhat better prognosis (8, 63). 150 Diabetes Mellitus Unfortunately, there have been few large series of patients with idiopathic or phenformin- associated lactic acidosis which allow evaluation of therapy. The effectiveness of any thera- peutic regimen is difficult to evaluate because of differences in diagnostic criteria and therefore the severity of the acidosis. In most instances, the requirement for bicarbonate is high and use of large quantities cannot reverse the acidosis. However, the early aggressive use of sodium bicarbonate with careful monitoring of pH may be more advantageous than the use of similar doses of bicarbonate over longer periods, recognizing that rapid alkalinization can decrease oxygen delivery to tissues and intensify lactic production (7). As a result of the resistance to bi- carbonate, indicative of continuing overproduction and/or underutilization of lactic acid, problems relating to sodium and volume overload are superimposed upon and compounded by the deleterious effects of continuing acidosis on cardiovascular function. Although dialysis is not an effective means of correcting the acidosis, it offers a mechanism for control of volume and body sodium content and therefore may be an important therapeutic adjunct (17). While there is no evidence on this point, dialysis may also be useful in the treatment of lactic acidosis in patients in whom phenformin is thought to play an etiologic role since removal of phenformin by this route may be of benefit. In extreme cases, methylene blue, an oxidizing agent which accepts protons and alters cellular redox state, can be used (64), however, in most cases it has not been effective (38). When lactic acidosis occurs with hypoglycemia the infusion of glucose and correction of the hypoglycemia has been associated with correction of the acidosis (37). Rare patients with diabetes mellitus and lactic acidosis have also had a beneficial response when insulin was used to treat accompanying hyperglycemia (29). In a recent review Dembo (16) observed that the recovery of patients with phenformin associated lactic acidosis who were treated with insulin was higher (65 percent) than those who were not so treated. RECAPITULATION The pathogenesis of idiopathic lactic acidosis and its relationship to diabetes mellitus is incompletely understood. Additional epidemiologic and basic information must be obtained before this relationship can be clarified. High priority must be given to studies which establish the true frequency of idiopathic lactic acidosis in acutely and chronically ill, hospitalized and ambulatory, diabetic and nondiabetic populations to determine whether the apparent predisposition of the diabetic patient to lactic acidosis is more apparent than real. Basic studies must also be undertaken to more thoroughly understand the factors which regulate lactic acid production and lactate homeostasis in animals and man before the mechanisms responsible for the development of idiopathic lactic acidosis and the role of diabetes can be fully defined. The importance of phenformin in this syndrome, particulary if it is to be retained as a therapeutic agent, will require continued investigation into its mechanism of action. The role of other drugs, such as ethanol and of drug-drug interactions in the pathogenesis of idiopathic lactate acidosis must also be considered in future research. Acknowledgments: The author would like to thank Dr. A. Vongries and the CIBA-GEIGY Corporation for allowing him access to previously unpublished data concerning phenformin and lactic acidosis. This work was supported by the Veterans Administration Hospital, Birmingham, Alabama, and by U.S. Public Health Service Grant, AM09722, and by a grant (RR-32) from the General Clinical Research Centers Program of the Division of Research Resources, National Institutes of Health. Lactic Acidosis 151 REFERENCES 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. Alberti, KGMM and TDR Hockaday 1972. Blood lactic and pyruvic acids in diabetic coma. Diabetes 21:350-368. Alberti, KGMM 1975. Personal communication. Altschuld, RA, and FA Kruger 1968. Inhibition of hepatic gluconeogenesis in guinea pigs by phenformin. Ann NY Acad Sci 148:612-617. Anderson, J, WD Thomas and RWS Tomlinson 1966. Pyruvate and lactate excretion in diabetes mellitus after sodium lactate. Brit Med J 2:114-121. Appel, KE, and DA Cooper 1927. Diabetic acidosis with a negative ferric-chloride reaction in the urine: Report of five cases. Am J Med Sci 173:201-220. Beckman, R 1968. The fate of biguanide in man. Ann NY Acad Sci 148:320-322. Bellingham, AJ, JC Detter and C Lenfant 1970. The role of hemoglobin affinity for oxygen and red cell 2,3-diphosphoglycerate in the management of ketoacidosis. Trans Assoc Amer Phys 83:113-120. Bengtsson, K, B Karlberg and S Lindgren 1972. Lactic acidosis in phenformin treated diabetes. Acta Med Scand 191:203-208. Berry, MN 1967. The liver and lactic acidosis. Proc Roy Soc Med 60:1260-1262. Bernier, GM, M Miller and S Springate 1963. Lactic acidosis and phenformin hydrochloride. JAMA 184:43-46. Cahill, GF, Jr., and OE Owen 1968. Carbohydrate metabolism and its disorders. London, Academic Press, pp 497-522. Cleaver, T, and R Caretta 1972. Lactic acidosis with phenformin therapy. Calif Med 117: 14-19. Craig, JW, M Miller, H Woodward, et al. 1960. Influence of phenethylbiguanide on lactic, pyruvic and citric acid in diabetic patients. Diabetes 9:189-193. Daughaday, WH, RJ Lipicky and DC Rasinski 1962. Lactic acidosis as a cause of nonketotic acidosis in diabetic patients. New Eng J Med 267:1010-1014. Davidson, MB, WR Bozarth, DR Challoner, et al. 1966. Phenformin, hypoglycemia and lactic acidosis. New Eng J Med 275:886-888. Dembo, AJ, EB Marliss, ML Halperin 1975. Insulin therapy in phenformin-associated lactic acidosis. Diabetes 24:28-35. Ewy, GA, RC Pabico, JF Maher, et al. 1963. Lactate acidosis associated with phenformin therapy and localized tissue hypoxia. Ann Int Med 59:828-833. Exton JH, LE Mallett, LS Jefferson, et al. 1970. The hormonal control hepatic gluconeo- genesis. Rec Prog Horm Res 26:411-462. Fajans, S, J Moorhouse, H Dooenbos, et al. 1960. Metabolic effects of phenethylbiguanide in normal subjects and in diabetic patients. Diabetes 9:194-204. Felig, P, and J Wahren 1971. Influence of endogenous insulin secretion on splanchnic glucose and amino acid metabolism in man. J Clin Invest 50:1702-1711. Freyschuss U, and T Strandell 1967. Limb circulation during arm and leg exercise in the supine position. J Appl Physiol 23:163-170. 152 22. 23, 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37- 38. 30. 40. 41. 42. 43. 44, Diabetes Mellitus Garland, PB, EA Newsholme, and PJ Randle 1964. Regulation of glucose uptake by muscle. A. Effects of fatty acids and ketone bodies and of alloxan-diabetes and starvation on pyruvate metabolism and on lactate/pyruvate and L-glycerol 3-phosphate/dihydroxyacetone phosphate concentration ratios in rat heart and rat diaphragm muscles. Biochem J 93:665-678. Gutler, F , B Peterson, and K Kjeldsen 1963. The influence of phenformin on blood lactic acid in normal and diabetic subjects during exercise. Diabetes 12:420-428. Hartmann, AF, and MJ Senn 1932. Studies in the metabolism of sodium r-lactate. Response of human subjects with liver damage, disturbed water and mineral balance, and renal in- sufficiency to the intravenous injection of sodium r-lactate. J Clin Invest 11:345-355. Hall, H, G Ramachander, and JM Glassman 1968. Tissue distribution and excretion of phenformin in normal and diabetic animals. Ann NY Acad Sci 148:601-611. Hockaday, TDR, KGMM Alberti 1972. Diabetic coma. Clin in Endocr and Metabo 1:751-788. Huckabee, WE 1958. Relationship of pyruvate and lactate during anaerobic metabolism I: Effects of infusion of pyruvate on glucose and of hyperventilation. J Clin Invest 37:244-254. Huckabee, WE 1961. Abnormal resting blood lacate I: The significance of hyperlactatemia in hospitalized patients. Amer J Med 30:833-839. Johnson, HK, and C Waterhouse 1968. Lactic acidosis and phenformin. Arch Int Med 122:367- 370. Jorfeldt, L 1970. Metabolism of L(+)-lactate in human skeletal muscle during exercise. Acta Physiol Scand Suppl 338. Kreisberg, RA 1968. Glucose metabolism in normal and obese subjects: Effects of phenformin. Diabetes 17:481-488. Kreisberg, RA, LF Pennington, and BR Boshell 1970. Lactate turnover and gluconeogenesis in obesity: Effect of phenformin. Diabetes 19:64-69. Kreisberg, RA, WC Owen, and AM Siegel 1971. Ethanol induced hyperlactatemia: Inhibition of lactate utilization. J Clin Invest 50:166-174. Kreisberg, RA, WC Owen, and AM Siegel 1972. Hyperlacticacidemia in man: Ethanol phenformin synergism. J Clin Endocr 34:29-35. Lacher, J, and L LaSagna 1966. Phenformin and lactic acidosis. Clin Pharm Therapy 7:477-481. Lexow, P 1959. Diabetic coma without ketosis. Acta Med Scand 163:115-119. Medalle, R, and C Waterhouse 1971. Lactic acidosis and associated hypoglycemia. Arch Int Med 128:273-278. Oliva, PB 1970. Lactic acidosis. Amer J Med 48:209-225. Olsen, RE 1963. "Excess Lactate'" in anaerobiosis. Ann Int Med 59:960-972. Owen, OE, AP Morgan, HG Kemp, JM Sullivan, MG Herrera, and GF Cahill, Jr. 1967. Brain metabolism during fasting. J Clin Invest 46:1589-1595. Patrick, SJ 1966. Effects of phenformin and hypoglycin on gluconeogenesis of rat tissue. Can J Biochem 44:27-33. Peretz, DI, M McGregor, and JB Dossetor 1964. Lactic acidosis: A clinically significant aspect of shock. Can Med Assn J 90:673-688. Proctor, DW 1967. Fatal lactic acidosis after an overdose of phenformin. Brit Med J 4:216. Rosenbloom, JA 1915. A form of diabetic coma not due to acetone bodies. NY Med J 102:294-296. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. Lactic ACidosis 153 Rowell, LB, KK Kraning, TO Evans, et al. Splanchnic removal of lactate and pyruvate during prolonged exercise in man. J Appl Physiol 21:1773-1783. Ruggles, TN, MH Lavietes, M Miller, et al. 1968. Effect of phenformin on the elevated blood lactic acid produced by hypoxia in normal and diabetic rats. Ann NY Acad Sci 148:662-670. Searle, GL, et al. 1966. Body glucose kinetics thylbiguanide. Diabetes 15:173-178. in nondiabetic human subjects after phene- Searle, GL, MD Siperstein 1975. Lactic acidosis associated with phenformin therapy: Evidence that inhibited lactate oxidation is causative factor. Diabetes (in press). Shepardson, CR, TG Christopher, M Miller, et al. blood lactate levels following exercise. J Lab Shreeve, WW, RC DeMeutter, and Y Shingeta 1964. 1962. The effect of phenethylbiguanide on Clin Med 60:1018. Diabetes, insulin, tolbutamide and glucose load in the degradation of Cl" labeled lactate and pyruvate. Diabetes 13:615-621. Steiner, DF, RH Williams 1958. Actions of phenethylbiguanide and related compounds. Diabetes 8:154-157. Sokal, JE, CU Lowe, EJ Sarcione, et al. 1961. Studies of glycogen metabolism in liver glycogen disease (Von Gierke's Disease): Six cases with similar metabolic abnormalities and response to glucogen. J Clin Invest 40:364-374. Sriussadaporn, S, JN Cohn 1968. Regional lactate metabolism in clinical and experimental shock. Circulation (Suppl 6) 37:181. Stainsby, WN, HG Welch 1966. Lactate metabolism Amer J Physiol 211:177-183, Strangaard, S, PE Nielson, V Bitsch, etal. 1971. of contracting dog skeletal muscle in situ. . Blood lactate and ketone bodies in diabetic ketoacidosis. Acta Med Scand 190:17-20. Tashkin, DP, PJ Goldstein and DH Simmons 1972. Hepatic lactate uptake during decreased liver perfusion and hypoxemia. Amer J Physiol 223:968-974. Tolstoi, E, RO Loebel, SZ Levine, et al. 1923-24. Lactic acids in diabetes following insulin. Proc Soc Exp Biol Med 21:449-452. Toews, CJ, JJ Cannon, JL Kyner, et al. 1970. The effect of phenformin on gluconeogenesis in isolated perfused rat liver. Diabetes 19:368. Tranquada, RE, S Bernstein, and HE Martin 1963. phenformin therapy. JAMA 184:37-42. Irreversible lactic acidosis associated with Traquada, RE 1964. Lactic acidosis. Calif Med 101:450-461. Traquada, RE, S Bernstein, WJ Grant 1964. Intravenous methylene blue in the therapy of lactic acidosis. Arch Int Med 114:13-25. Traquada, RE, WJ Grant, CR Peterson 1966. Lactic acidosis. Arch Int Med 117:192-202. Vongries, AG 1973. Personal communication. Wahren, J, P Felig 1971. Glucose metabolism during leg exercise in man. J Clin Invest 50: 2715-2725. Walker, RS, AL Linton, WST Thomason 1960. Mode hydrochloride. Brit Med J 2:1567:1569. of action and side effects of phenformin Waters, WE, JD Hall, WB Schwartz 1963. Spontaneous lactic acidosis. The nature of the acid-base disturbance and considerations in diagnosis and management. Amer J Med 35:781-793. Watkins, PJ, JS Smith, MG Fitzgerald, et al. Lactic acidosis in diabetes. Brit Med J 1:744-747. 13 ACUTE COMPLICATIONS OF THE DIABETIC STATE Joseph Silva and F. Robert Fekety, Jr. INFECTION The Frequency and Importance of Infection in Diabetes Mellitus It is generally believed that diabetics have more infections than nondiabetic persons, and that infections in diabetics tend to be more severe and difficult to manage (62,67). However, this is still controversial and some believe the relationship has been overemphasized. While we believe that anyone who treats diabetics will soon be convinced of the inordinate frequency, devastating effects, and great importance of infections, the arguments on both sides should be presented and put into perspective. Unfortunately, little scientific evidence exists to prove that diabetics have a generally increased rate of infection and a clinically significant impair- ment of resistance to most infectious agents. There is even some evidence that the increased frequency of certain infections in diabetes is attributable not to the diabetic state per se but instead to harmful things that happen to diabetic patients. This line of reasoning is attractive because some of these things may be eliminated. For example, it is maintained that the high in- cidence of urinary tract infections in diabetics is attributable to the frequency of bladder cathetrization during acidosis or to relieve a distended neurogenic bladder. Similarly, frequent skin infections may be attributed not to difficulty in handling microorganisms, but instead to obesity and poor personal hygine, to insulin injections using unsterile equipment or to the in- ability to care for oneself because of diabetic retinopathy or cataracts. Infections of feet and legs may be attributed to arteriosclerosis and vascular insufficiency instead of to diabetes. Un- questionably, these factors are important, but they incompletely explain all these infections. More importantly, some infections are clearly more common in diabetics even in the absence of such factors. Most investigators conceded that Candida infections of the perineum and vagina, and Phycomycotic infections of the orbit and brain are more common in the presence of diabetes. Another erroneous concept that needs to be dispelled is that infectious diseases have become infrequent and relatively unimportant both in normals and in diabetics. The reasons proposed for this include the ready availability of potent antibiotics and vaccines, and an increasingly good standard of living associated with a decline of contagious diseases thriving on poverty, bad hous- ing, and lack of sanitation. For example, Younger (67) has pointed out that with insulin therapy, but prior to the discovery of antibiotics, the average duration of life in juvenile diabetics was about seven years, and that death was usually related to sepsis. With insulin plus antimicrobial chemotherapy, the prognosis is better than two-thirds of normal life expectancy. However, it should be emphasized that this represents a major shortening of life. In addition, she goes on to point out that although the prognosis for diabetics as a group is now favorable, for many in- dividuals within the group it is devastatingly unsatisfactory. Younger maintains that micro- angiopathy is the major factor associated with a bad prognosis and that infection is an important contributory factor to angiopathy. The role of infection in aggravating microangiopathy is not well established or clearly understood and needs further study. 154 Acute Complications 155 In another related study, Robbins and Tucker (53) noted that when the causes of death of 307 diabetic patients over the age of 12 were compared with those of 2800 consecutive nondia- betic patients in 1944, the relative incidence of pulmonary and other infections found at autopsy was approximately the same in both groups. One major defect in this study is that the patients were apparently not matched according to age, race, and sex, and the data really prove little more than that infections are frequent in people who die from any cause. Another defect is that the data do not touch upon the tremendous morbidity and suffering from nonfatal infections that diabetics must experience and endure. Nonetheless, two relatively common and serious infections, acute pyelonephritis and infections of the extremities were observed more frequently among dia- betics in this study. The concept that infections are not important in diabetics is invalid. Two recent studies merit attention because they vividly document how infections are still of great significance to patients with diabetes. The first concerns patients who were admitted in diabetic ketoacidosis to the Birmingham General Hospital in the United Kingdom between January 1968 and October 1972. The patients were treated in an intensive care unit by a team skilled in the management of diabetic acidosis (59). The mortality rate in 258 episodes of diabetic keto- acidosis was 6 percent. Infection was responsible for ketoacidosis in 78 (38 percent) of the 190 episodes in which a factor precipitating the episode could be identified. The next most com- mon precipitating factor, error in insulin dosage, was only half as frequent as infection. Seven (10 percent) of the 73 patients with infection died, while only 5 percent of those with keto- acidosis, not complicated by infection, died. Infection accounted for 7 (43 percent) of the 16 deaths in this study. One of the fatal cases had meningitis, three had influenza, and three had pneumonia. The mean age of the patients in this report was only 41 years, emphasizing the im- portance of infection as a serious complication even in relatively young diabetics. Similarly, Muller and his colleagues (44) reported on 26 consecutively admitted patients with diabetic keto- acidosis in Dallas during 1969 and 1970, and indicated that 20 (77 percent) of the 26 had infec- tion as the cause of diabetic ketoacidosis. A wide variety of common and mundane infections that are supposedly no longer important were implicated in their patients. What should be emphasized in assessing the significance of these recent reports is that re- search in the infectious diseases that harrass and kill diabetics has not been of major interest to most physicians and funding agencies concerned with endocrinology and metabolism. Quite under- standably, they have been more concerned with important biochemical and cardiovascular features of the diabetic state. Therefore, in the future, research funding agencies interested in allevi- ating suffering in diabetic patients ought to take into consideration the present scarcity of funds for research in infectious diseases. Recently this has become increasingly dependent upon pharmaceutical companies. Furthermore, it should be stressed that research in the broad field of infectious diseases is likely to yield significant benefits to diabetic patients, because they suffer most from the common infections afflicting everyone. Looking to the future, it is likely that infections will be even more serious and costly. For example, Kjellstrand and his colleagues (33) noted that uremia is a common cause of death of .patients with juvenile-onset insulin-dependent diabetes, and that renal transplants are being performed in them with increasing frequency. In reporting their experience with 50 patients, they stated that the overall results were encouraging and suggested many of the symptoms of ad- vanced diabetes may be caused by uremia rather than the diabetic process. On the other hand, 15 156 Diabetes Mellitus of their 40 patients died, and the main cause of death was infection. Thus, while it is clear that effective antibiotic treatment of infection is second only to insulin in increasing the life span of a diabetic patient, infections cannot be discounted now or in the foreseeable future as a major cause of morbidity and mortality. They should be a major subject of further research, which should not be left to the marketplace needs of the pharmaceu- tical industry. In order to emphasize areas of special importance for future research, we will discuss the more noteworthy specific infections. Infectious Processes Associated with the Onset of Diabetes More and more attention is being given to the probability that the diabetic state may be caused, precipitated, or accelerated by viral or other infectious diseases causing inflammation of the pancreas. We believe this is an important area for further research, which is mentioned only in passing because it is discussed in detail in another section. Skin and Soft Tissue Infections It is probable that staphylococcal skin infections are more common in diabetic patients. This predilection may be related to their frequent contact with antibiotics and hospital reser- voirs of resistant staphylococci, and to the apparent increased rate of nasal carriage of staphylococci in diabetics. It may also relate to the use of unsterile needles and poor personal hygiene. Generalized hypersusceptibility to staphylococci in diabetes is not proven, and dia- betes does not appear important in increasing the rate of staphylococcal post-operative wound in- fections (12). Diabetes is frequently implicated in staphylococci bacteremia requiring hospital- ization (11). Further definition of the responsible mechanism(s) is needed if prevention is to be made possible. Staphylococcal infections are especially worthy of further study because of the great frequency and severity of these infections in diabetes. Peripheral vascular disease is a major factor contributing to staphylococcal and streptococ- cal infection of the legs in the diabetic. These organisms are particularly important causes of ulcers, gangrene, osteomyelitis, and other infections of the feet and legs that ultimately may result in amputations. Antimicrobial as well as surgical therapy of these infections must be im- proved if extremities are to be saved and mobility and independence preserved. Minor infections of the extremities often become serious and life-threatening to diabetics if treated improperly in early stages. Prevention of infection by means of optimal foot care and avoidance of burns and injuries in patients with diabetic neuropathy is most important. Candidiasis (Moniliasis) is a yeast (fungal) infection commonly afflicting diabetics. It results in marked redness and swelling of the genitalia, perianal region, medial aspects of the thighs and other moist, intertriginous areas, Small superficial pustules are usually present at the edges of the involved areas. Itching and soreness are usually severe. The process may involve the oral mucous membranes (thrush) and occasionally becomes generalized. Candidiasis is common in patients with poorly controlled diabetes, and is often the initial symptom of the dis- ease. It is well recognized that all patients suffering from Candidiasis should be tested for diabetes. When localized to the skin, Candidiasis is relatively minor and easily controlled, especially if the diabetes is well-managed, but occasionally these infections become generalized and systemic in seriously ill, hospitalized diabetic patients. These disseminated infections are extremely refractory to treatment and are associated with a high mortality rate. Acute Complications 157 Pneumonia There is little published data on pneumonia in diabetics, even though most clinicians believe these are serious illnesses for them. In a study of 62 patients admitted to the Johns Hopkins Hospital in 1965-1966 because of pneumococcal pneumonia, diabetes was evident in only four patients, and was no more frequent than it was in a carefully matched control population (23). Nonetheless, another recent publication reporting on 112 diabetics with pneumonia called atten- tion to the impact of this infection (31). The mortality from pneumonia in this study was 39 percent and the frequency of prolonged infection with such dangerous pathogens as Staphylococci and Klebsiella was high. This calls attention to the problem of hospital-derived pulmonary in- fections in diabetics. These probably are increasing in frequency, and better ways of preventing and managing these problems are needed. Tuberculosis Tuberculosis has long been a serious problem in association with diabetes mellitus, but the availability of effective antituberculous chemotherapy has significantly improved the prognosis for this disease. The physician's emphasis now should be upon adequate detection of early tuber- culous infections by means of annual tuberculin skin testing and periodic chest x-rays. This will result in early treatment, with a high probability of success. Guidelines for managing dia- betics who develop tuberculosis are well established and diabetic patients usually respond quite satisfactorily to appropriate chemotherapy. Tuberculosis should no longer be a major threat to the life of patients with diabetes. This is an outstanding example of how progress in a distant area can have profound effects upon diabetes. Urinary Tract Infections Catheterization, neuropathy with bladder paralysis, nephrosclerosis, and Kimmelsteil-Wilson disease contribute to the problem of urinary tract infections in diabetics. This infection com- monly results in pyelonephritis and may eventually progress to renal failure. Urinary hyper- osmolarity attributable in part to the excretion of glucose may result in impaired leukocyte function in the medulla of the kidney (10) and may be of importance in the pathogenesis of pyelonephritis and necrotizing renal papillitis. Necrotizing papillitis ascribed to local ischemia and accompanied by septicemia was demon- strated in 10 of 266 post-mortem examinations of diabetic patients who died during a 5-year period (66). Obstructive uropathy is important in its pathogenesis, as is the use of analgesics such as phenacetin and possibly aspirin. The clinical picture is not always fulminating, and papillitis should be entertained in every diabetic patient with infection and worsening renal function. Gellman and his associates (25) found chronic pyelonephritis in approximately 10 percent of 63 renal biopsies in diabetic patients. The incidence of pyelonephritis was even higher in autopsied cases, and in some series has ranged as high as 40 to 55 percent (67). The disease seems more frequent in females than males, a fact which is probably related to the short urethra of women, to pregnancy, and to the high frequency with which women are catheterized. The frequency of asymptomatic bacteriuria in diabetics is also apparently higher than in matched nondiabetic patients, but the statistics vary widely from place to place. Asymptomatic bacteruria is thought to be an important but not inevitable precursor of chronic pyelonephritis 158 Diabetes Mellitus and uremia. Indwelling urethral catheters with their inherent risk of asymptomatic or overt urinary tract infection and sepsis are important causes of serious morbidity and mortality from infections in diabetics. Many already well established measures and concepts are influential in reducing the frequency of this complication, but they need further refinement and development to insure popularity, practicality, and wide application. For example, the value of using closed sterile drainage systems or continuous bladder irrigation with antibacterial agents in patients with indwelling catheters needs to be emphasized. There is a need for further research and de- velopment of new preventive measures in this important area. Gram-Negative Bacteremia and Sepsis It is difficult to substantiate the belief that Gram-negative bacteremia is more common in diabetics, and that shock and mortality rates are increased. Diabetes was found in 18 (20 per- cent) of 88 patients with Gram-negative sepsis at the Johns Hopkins Hospital (35), and the case fatality rate for diabetics (39 percent) was higher than for the entire group of septic patients (24 percent). The special hazard of these infections probably is related at least in part to exposure to hospitals and their attendant hazards such as intravenous and urinary catheters, to unconsciousness with aspiration, and to inhalation therapy with the increased risk of development of pneumonia (62,67). While a poorly explained phenomenon, it is well documented that persons over 55 with serious underlying diseases such as diabetes have a higher than average mortality rate from Gram-negative sepsis, and this condition is thus an important problem for further re- search in pathogenesis and treatment. Phycomycosis (Mucormycosis) Phycomycosis is a rare fungus infection that is caused by saprophytic organisms of the genus Phycomyces (including Rhizopus, Mucor, and closely related species frequently found on fruits and vegetables). It occurs most often in patients who are acidotic or receiving adrenal steroid therapy and is associated with diabetes mellitus in the majority of instances. After colonizing the nasal passages, these organisms sometimes invade the adjacent tissues, thrombosing blood vessels and producing necrosis and gangrene of the tissues of the nose and eye. Cerebral exten- sion of the process occurs in about two-thirds of the cases, producing paralysis of eye muscles and signs of diffuse cerebrovascular disease and is usually fatal. The condition seems to be re- lated to ketosis more than to acidosis, and has been correlated experimentally with a brief delay in leukocyte mobilization at the site of primary invasion by the fungus (55). This delay permits a brief period of unchecked proliferation of the organism and is probably illustrative of an im- portant defect which is operative in many other serious infections in diabetes. Mucormycosis is treatable with amphotericin B, but it has a high fatality rate, and extensive and disfiguring surgery to remove necrotic and gangrenous tissues is often required. Blindness in the involved eye is usual, and prolonged and expensive hospitalization is the rule. Host Factors Related to Hypersusceptibility to Infection in the Diabetic While it is still unproven that diabetics have a general increase in susceptibility to in- fection, there can be no doubt that infections are of tremendous importance to them. Because a wide diversity of organisms are involved at different body sites, the conclusion is inescapable that deficiencies in many of the cellular and humoral immune systems are probably implicated. Furthermore, these immune mechanisms may be deranged in different ways for different infectious Acute Complications 159 diseases. Many studies designed to elucidate these abnormalities have yielded conflicting and inconclusive results. In most cases, the mechanisms have not been established, and although many interesting leads have been developed, much further work is needed. As a probable first step in elucidating these mechanisms, prospective studies of the rates of various types of infection in diabetic populations and matched controls should be performed. Those infections shown to be clearly more frequent can then be studied in more depth to determine the responsible factors. Factors that already are considered potentially important in the pathogenesis of infection in diabetes will be presented according to the following outline: (a) factors resulting in increased exposure to microorganisms, (b) factors facilitating entry of microorganisms into susceptibel diabetic tissues, (c) specific defects in cellular and humoral defense mechanisms in diabetes, and (d) metabolic abnormalities which may nonspecifically contribute to hypersus- ceptibility. 1. Factors increasing exposure to organisms. Healthy human beings harbor large numbers of bacteria and fungi on the skin, and within the mouth, nose, pharynx, and gastrointestinal tract. This is the so-called normal or resident flora, which may be found at these sites in numbers which vary from 107 to 101! organisms per gram or square centimeter of tissue sampled. These organisms are in a continual competitive struggle with one another. Antibiotics and certain underlying diseases can alter the balance between species, and despite their weak pathogenicity, disease may result if the numbers of organisms markedly increases. Alternations in the composition of resident flora are seen in some diabetics, but the reasons for them are unknown. Samplings of diabetic skin, stool, and mouth flora have demonstrated an increased rate of carriage of Candida (29) and of staphylococci in the nose (57). There is little information concerning whether the adequacy of diabetic control influences the quantitative load of bacteria or fungi at various sites. Another factor influencing the acquisition of pathogenic bacteria is that diabetics have frequent contact with hospitals, clinics, and health personnel. These locations and individuals may serve as reservoirs of microorganisms that are more pathogenic and resistant to antibiotics. For instance, the acquisition of Staphylococcus aureus infections which are resistant to penicil- lin is much more frequent in hospitalized patients than in outpatients. Infection with such organisms may be far more serious than with an organism acquired at home. 2. Factors facilitating the entry of microorganisms into susceptible tissues. The diabetic patient may also be compromised in defending against infection because vascular and neurological complications of the diabetic state permit organisms to gain a foothold in normally inaccessible tissues. For instance, ketoacidosis leads to coma and dehydration, which decreases the produc- tion of oropharyngeal and tracheal secretions. These secretions are important in providing a mucous blanket which defends against penetration of bacteria and also washes them away. Not surprisingly, pneumonia frequently occurs during coma because of aspiration of organisms and failure to clear them. The neurological complications of diabetes include a variety of neuropathies. A motor or autonomic neuropathy may lead to incomplete drainage of the bladder, which causes stasis and urinary retention requiring catheterization with occasional introduction of organisms and subse- quent urinary infection. Autonomic neuropathies of the gastrointestinal tract may produce a malabsorption syndrome. Vitamin deficiencies and malnutrition may then occur, which further predispose to recurrent infections in poorly understood ways. Sensory neuropathies cause sensory 160 Diabetes Mellitus deprivation leading to neglect of wounds, contusions, abrasions, and lacerations, which are sus- ceptible to bacterial invasion. } The vascular diseases which are so frequent in diabetes may produce varying degrees of ischemia or gangrene in peripheral tissues. The frequency of gangrene in diabetes is well-known and much feared by laymen, for gangrene predisposes to extensive infection of extremities and bacteremia and may result in amputation or death. Organs or skin deprived of a vascular supply are much more prone to infection. Local defenses are compromised because of poor delivery of exudative substances such as antibodies and inflammatory cells, and of antibiotics. Thus, these vascular or nervous system complications subject the diabetic patient to the hazards of urinary and intravenous catheterizations for drainage or for the delivery of fluids, electrolytes, and insulin. Indwelling intravenous catheters are being associated with infections at an increasing frequency. These may vary from focal phlebitis to bacteremia and septic shock. The responsible organisms are usually derived from the patient's skin, and may gain entry to the tissues via the penetrating catheter which may be left in place for long periods. Urinary catheters become colonized within a week of placement and can serve subsequently as a focus re- sulting in pyelonephritis or bacteremia. Thus, the poorly managed ketoacidotic diabetic patient requiring various catheters is subjected to an increased risk of infections that he is poorly able to withstand. Furthermore, when faulty techniques are used, abscesses or cellulitis leading to bacteremia can occur at sites of insulin injection. 3. Specific defects in cellular or humoral defense mechanisms. The inflammatory response has three major components: (a) antibodies and other chemical substances such as histamine, complement, and prostaglandins, (b) phagocytes (granulocytes, macrophages), and (c) lymphocytes. The first two components dominate the acute inflammatory reaction, whereas the latter component is especially involved in the more complicated cell-mediated response of a delayed type. Defense against pyogenic infections is usually dependent upon the first two components, whereas defense against some fungi and bacteria such as Mycobacterium tuberculosis is usually attributed to the delayed cellular immune response. These systems seem to overlap and may reinforce one another, especially if one component is defective. Some of the metabolic abnormalities that occur in the diabetic patient and may influence these immune systems include hyperglycemia, alterations in serum and tissue osmolarity, and in- creased serum levels of ketones, lactic acid, and lipids. Investigators have examined the in- fluence of these variables on some aspects of inflammation, mainly in vitro. There are only a few in vivo studies of immune responses in diabetic patients. The resultant information concern- ing inflammatory responses related to hypersusceptibility in the diabetic is sparse and often contradictory. Alterations in Antibody Mechanisms in Diabetes. Antibodies are gamma globulins which may be natural" or nonspecific, while others are produced against specific invading organisms. These antibodies kill organisms in a variety of ways, such as by direct lysis (bactericidal antibodies), by enhancing bacterial ingestion by phagocytes (opsonizing antibodies), or by agglutinating or precipitating clumps of organisms, thus allowing more effective clearance by circulating and fixed phagocytes. Many of these reactions are enhanced by low molecular weight proteins which comprise the complement systems (including properdin). Data concerning antibody deficiencies in diabetic animals and patients are conflicting and inconclusive. Studies on the ability of diabetic patients to form antibodies to a variety of antigens have Acute Complications 161 indicated either an impairment (17,52) or no abnormalities (38,63). Studies of gamma globulin levels (36) and the properdin system (27) showed no abnormalities in the well-controlled diabetic patient. Powell and Field (51) and Balch et al. (6) found diabetic patients had increased levels of serum complement, indicating that these circulating proteins are probably not defective in diabetics. Because of antibodies and other substances, blood from normal persons is bactericidal for many organisms. The addition of glucose (up to 10 times normal serum concentration) to blood from normals, does not decrease its bactericidal properties, nor increase the rate of growth of most microorganisms in culture media (50). In contrast, Richardson (52) found that the growth of several types of bacteria in artificial media was increased in the presence of whole blood from diabetic patients but not in the presence of normal blood. The mechanism of this effect is un- known and deserves further study. Other factors besides deficiencies in antibodies may alter the bactericidal activity of blood. Dubos (19) showed that keto-acids protected bacteria from the natural bactericidal action of lactic acids. This may relate to the clinical observation that diabetics are especially prone to infection when ketoacidosis prevails. Studies utilizing experimental models of diabetes sug- gest that the serum bactericidal capacity may be deficient in some circumstances. Cruickshank and Payne (15) showed that blood of alloxan diabetic rabbits had a modest decrease in bacteri- cidal capability for pneumococci of low virulence, but further experiments using more virulent pneumococci, staphylococci, and M. tuberculosis showed no abnormalities (14). Balch et al. (6) in a well controlled study of diabetic patients (with an without infection) found no alterations in serum bactericidal activity which correlated with glucose or ketone levels. Furthermore, these patients showed no consistent depression of serum bactericidal ac- tivity when their insulin was withheld for 48 to 96 hours. These data suggest that blood from most diabetic patients has a normal bactericidal capacity. More sensitive methods of studying antibody formation and function than were used in these studies are now available and should be applied to the question of whether antibody responses are defec- tive in diabetics. Deficiencies in Phagocytic Metabolism. Investigations of leukocyte morphology in diabetes have indicated few defects in clinical importance. The total numbers of blood leukocytes are not abnormally low in diabetes. In fact, leukocytosis and basophilia are common. Phagocytes have a variety of functions that can be studied in vitro: (a) random migration, (b) specific and nonspecific chemotaxis (purposeful migration) which can be measured in chambers or in the skin, and (c) ingestion and digestion of bacteria and fungi. In addition, metabolic events such as the rates of hydrogen peroxide generation and oxygen and glucose consumption can be studied in leukocytes. Certain enzymes such as nitroblue tetrazolium (NBT) reductase have been studied because of their potential importance in the intracellular killing of bacteria. Normal phagocytes (including polymorphonuclear leukocytes and monocytes) metabolize glucose via the Krebs-Meyerhof pathway (anaerobic glycolysis) in order to migrate and ingest orgnisms. Glucose metabolism shifts to aerobic pathways following ingestion of organisms or particles. This shift in glucose metabolism seems important, as bactericidal complexes are generated intra- cellularly in the process, and phagocytes lose their ability to kill bacteria when placed in anaerobic environments. During this process, oxygen is consumed and hydrogen peroxide is pro- duced. The hydrogen peroxide (H O ) then combines with a halide and an enzyme (myeloperoxidase) 2 2 162 Diabetes Mellitus to form a powerful bactericidal complex Some metabolic observations relevant to this system have been made utilizing diabetic phagocytes. Martin et al. (40) found decreased glucose utilization and lactate production in leukocytes obtained from diabetic patients in good control. Dumm (20) also found that leukocytes obtained from nonacidotic diabetic patients exhibited a depression of glycolysis which could be corrected by insulin. This is interesting because Esmann (22) showed that the phagocytic membrane does not require insulin for glucose transport. Consequently, functional defects in diabetic phagocytes, if attributable to an insulin deficiency, probably involves insulin's intracellular actions rather than a membrane effect. Insulin is known to regulate several rate-limiting glycolytic enzymes such as phosphofructokinase, pyruvokinase, and glucokinase (65). The oxidative metabolism of leukocytes from diabetic patients has been analyzed by direct oxygen measurements and NBT reduction (64). Diabetic leukocytes were found to have a lower degree of NBT dye reduction and a higher oxygen consumption than control leukocytes. These abnormally low NBT tests are interesting because these diabetic patients also had a reduction in phagocytic activity. The NBT reduction test is similarly abnormal in children with the syndrome of chronic granulomatous disease. In this condition, phagocytes are unable to kill certain microorganisms and the patients usually die of infection at an early age. These studies in diabetics are impor- tant and should be confirmed. Fixed tissue macrophages in diabetes have been described as having increased glycogen and cholesterol (46), which may relate to a potentially significant abnormality in these phagocytic cells. The effects of the intracellular accumulation of lipid, although described in diabetics in 1925, has to our knowledge not been fully analyzed and may be an important aspect of a compro- mised reticulo-endothelial system. The proof of the importance of any leukocyte defect lies in the answers to two questions: (a) do leukocytes migrate into an area of microbial invasion normally, and (b) do phagocytes ingest and kill microorganisms normally at these sites? Deficiencies in Chemotaxis (Migration) by Phagocytes. Several studies have shown that chemotaxis is not altered when leukocytes from non-diabetic individuals are incubated in glucose concentrations varying from 200 to 1000 mg percent. However, Perillie, Nolan, and Finch (49) noted a delayed and diminished migratory response of granulocytes in diabetic patients, as measured by the skin window technique. This effect was noted when the diabetic patients were acidotic but not when well-controlled. Furthermore, the deficient mobilization of these im- portant defenses returned to normal following correction of the acidosis. In vitro chemotactic defects have been noted in phagocytes from adult (7,26,43)and juvenile diabetic patients (41). There was no correlation between the chemotactic defect and levels of plasma insulin, or serum glucose, carbon dioxide, or urea nitrogen. The addition of insulin to the cell suspension corrected these chemotactic defects in one study (43), but not in others (26,41). The reasons for these observed chemotactic defects are not known. Similar defects in chemotaxis have been found in patients with multiple myeloma and macroglobulinemia and have been attributed to abnormal amounts of surface immunoglobulins. However, Mowat and Baum (43) could not demonstrate any abnormal coating of diabetic leukocytes with immunoglobulin. Acute Complications 163 Some of the chemotactic defects demonstrated in human diabetes have been noted in experi- mental models of diabetes. Cruickshank (14) noted that the inflammatory response in the skin following challenge with staphylococci was diminished in alloxan-diabetic, ketotic rabbits. Similarly, Briscoe and Allison (8) showed that the formation of peritoneal exudates was reduced in experimental peritonitis in nonketotic diabetic rats. Rabbits developed a significant reduc- tion in the intensity of their inflammatory reactions following relatively short periods of glucose infusion (2). This abnormality could be dissociated from hyperosmolar or acidifying ef- fects of glucose infusions. Thus, the mechanism of these defects is unclear and should be further evaluated. Other exudative factors such as pH that may relate to chemotactic defects are also worthy of consideration. The pH of inflammatory exudates in diabetic rabbits is lower than in non-diabetic rabbits. Hutchins and Sheldon (28) have shown that the reduced concentration of hydrogen ion at sites of injury in diabetic rabbits is associated with profound alterations in the immune defense. Furthermore, there is a direct correlation between the development of mucormycosis in the rabbit and the delay and paucity of the inflammatory response during ketosis (55). Organisms were noted to grow and spread through tissues before an effective inflammatory response could be mounted. These defects are especially significant because many investigators believe that the initial de- fensive reactions determine whether an infection will be established following the lodgement of organisms in tissues, and that the issue is usually settled within the first 30 minutes of tissue invasion. Deficiencies in Phagocytic Ingestion and Digestion. The data are conflicting about whether defects occur in ingestion and digestion by phagocytes in diabetes mellitus. Bybee and Rogers (9) reported that leukocytes of ketoacidotic, diabetic patients demonstrated a reduction in phagocy- tosis in vitro. The defect disappeared when the acidosis was corrected. Ingestive capacity of the diabetic leukocyte was restored by suspending the cells in normal instead of diabetic serum. However, leukocytes from normal individuals failed to develop the abnormality after being sus- pended in sera from ketoacidotic diabetic patients. Similarly, Bagdade et al. (5) showed that phagocytes from poorly controlled diabetics demonstrated a combination of inadequate ingestion and bactericidal activity. Insulin treatment reversed these abnormalities, and phagocytic ef- ficiency was inversely correlated with the fasting glucose level. Bagdade et al. (4) subsequently compared a group of nondiabetic patients with eleven poorly controlled diabetic subjects studied before and after treatment. Leukocytes from these patients demonstrated a defect in killing type 25 pneumococci which was related to an impairment in ingestion of bacteria rather than to their bactericidal capacity. The degree of impairment also correlated closely with the fasting glucose level and was corrected by appropriate treatment of the hyperglycemia and ketoacidosis. These studies have important implications, as they suggest that insulin corrected a potentially serious defect in leukocyte function. However, the test organism used was relatively avirulent. There- fore, these studies should be repeated with more virulent organisms and in particular with those bacteria and fungi which commonly infect diabetics. Recently, Tan et al. (60) found that the neutrophils of a diabetic patient who was suffering from an infection with Staphilococcus aureus were defective in killing this organism. However, Crosby and Allison (13,41) found that the bactericidal and phagocytic activities of leukocytes from diabetics who did not have ketoacidosis were not different from those of control subjects. 164 Diabetes Mellitus Better controlled study of several different kinds of microorganisms using leukocytes from dia- betics in varying degrees of control are obviously needed to answer the question of whether clinically significant intrinsic defects of phagocytic function occur. Studies of leukocytes obtained from animals with induced diabetes have shown as variable results as those from humans. Drachman et al. (18) showed a defect in the ability of circulating and lung phagocytes obtained from diabetic, nonketotic rats to kill type 25 pneumococci. While suggesting this deficiency was due to an intrinsic leukocyte abnormality, they believed the alveolar macrophage defect was related to serum hyperosmolarity attributable to hyperglycemia. However, Briscoe and Allison (8) could not find a defect in phagocytosis in nonketotic diabetic rats. Using avirulent pneumococci, Cruickshank and Payne (15) showed a decrease in bactericidal capacity of phagocytes obtained from alloxan diabetic rabbits which was related to ketoacidosis. Thus, these data from experimental models of diabetes mellitus suggest that intraleukocytic phagocytic defects in killing bacteria may occur and be related to the severity of diabetic con- trol, 4. Nonspecific defects and metabolic abnormalities relating to infection. Many physicians have been attracted to the notion that hypersusceptibility in diabetes has been related to hyper- glycemia and a saturation of tissues with glucose, leading to a more favorable environment for growth of microorganisms. The factors which compromise immunity in diabetes seem more complex than just hyperglycemia. Supplementation of serum or bacteriological media with glucose does not enhance bacterial growth rates appreciably (50). Several parameters of the immune system are more dramatically affected by ketoacidosis than by hyperglycemia. The ill effects of the hyper- osmolarity which occurs in the ketoacidotic state may be the important factor. Sbarra et al. (54) and Allison and Lancaster (3) have shown the dramatic adverse effects of hyperosmolarity on phagocyte function. However, the ranges of hyperosmolarity studied exceed those usually noted in the blood stream and tissues, except for the renal medulla (33). This may be important because diabetics have an increased rate of chronic pyelonephritis, which may be related to altered leukocyte function in the renal medulla (10). Alterations in complement frac- tions may also occur in this region and may interfere with effective opsonization and ingestion of bacteria. Another potentially important factor is that the diabetic patient may develop renal failure and uremia because of vascular disease or pyelonephritis. Uremic patients seem more prone to infection, but the responsible mechanisms are obscure (42). Malnutrition secondary to a chronic catabolic state may be another important factor in rela- tion to susceptibility in diabetes. Recent studies of phagocytes from humans (56) and lymphocytic-mediated cellular immunity (58) have shown defects related to severe malnutrition. Possibly lesser degrees of malnutrition as occurs frequently in diabetics may also affect leuko- cyte function. The relationship between the pathogenesis of infections and alterations in serum lipids is as yet unexplored. Abnormalities in lipid and fatty acid composition and production seem important in the pathogenesis of cutaneous infections, such as boils and acne. A variety of infections, such as cholera and influenza, and administration of bacterial endotoxin are associated with characteristic lipid alterations (24). These have not been related to changes in lipids in dia- betes as yet, but recent work suggests lipids influence the immune response to Coxsackie virus in Acute Complications 165 mice (37). This is an area worthy of further study. Summary of Abnormalities in Host Defense Mechanisms We believe that there are definite pathologic mechanisms operating in the diabetic patient, which predispose to recurrent infections. There are studies to support the presence of defects in chemotaxis, mobilization, and possibly ingestive and digestive capacities of phagocytes in diabetes. The ready availability of potent antibiotics has dampened enthusiasm for investigative ef- forts on the immune defects of diabetics. There are almost no data concerning host defenses against viral infections in diabetics. Whether cellular immunity, delayed hypersensitivity, and other lymphocyte-mediated immune events are intact in diabetics is largely unexplored. The im- portance of the potential defects is underscored by the recent data showing that diabetic patients with ketoacidosis frequently die of infections with organisms in which these systems play an im- portant defensive role (59). Further investigations should be conducted on the pathogenesis of these infections in dia- betes. Thomas (61) has indicated that research directed at corrective problems ("half way technology") is enormously expensive and often fails to affect disease, but that research directed at an understanding of disease processes ("high technology') is often rewarded with successful disease management and gratifying long-term benefits. The Adverse Effects of Infections Upon Metabolic Processes in Diabetes Mellitus While it is recognized that worsening of the diabetic state can precipitate infections such as mucormycosis, the reverse is probably more frequent. That is, infection increases insulin re- quirements and is important in worsening the diabetic state and precipitating ketoacidosis. As mentioned earlier, in two recent studies infection was the most frequent and important factor precipitating episodes of ketoacidosis (44,59). The exact mechanism by which infection is dia- betogenic and aggravates the disease is unclear, but there are several interesting clues. In the first place, infection is stressful and associated with an increased activity of adrenal steroid and epinephrine-like substances, both of which have profound effects upon carbo- hydrate and fat metabolism and commonly cause hyperglycemia. Second, fever in infection in- creases the rate of metabolic processes and increases caloric and water requirements. Third, relatively high levels of circulating glucagon have been found in humans with dia- betes (1,48) or infection. Glucagon is an insulin-opposing hormone. Glucocorticoid treatment in man has been shown to result in hyperglucagonism (39), and the increased endogenous steroid activity associated with infection has already been noted. Muller, Faloona, and Unger (44) have presented evidence suggesting that infection is asso- ciated with hyperglucagonemia which may thus worsen the diabetic state. They studied 26 patients with diabetic ketoacidosis. Infection was the precipitating event in more than half of these. Plasma glucagon levels average 390 pg/ml, which was significantly greater than the fasting level of 118 pg/ml in diabetic subjects without ketoacidosis. Absolute hyperglucagonemia was present in 16 of the 26 patients. Insulin was present in the plasma in six of seven ketoacidotic patients; the level was 'mormal'" in four. These studies of Muller and his associates (44) may explain, at least in part, the remarkable degree of insulin resistance which almost universally characterizes the early hours of severe diabetic ketoacidosis. While severe insulin deficiency can cause hyperglucagonemia (45), they 166 Diabetes Mellitus believed absolute insulin lack was probably not responsible for the metabolic deterioration noted in most of the patients in their study, since four had normal levels of insulin during their acute illness. If their data and reasoning are correct, an infection-related stimulus caused an inap- propriate release of glucagon. The mechanism of this is as yet unknown and seems worthy of fur- ther study. Unger and his associates reported that dialysates of infected tissues stimulated glucagon secretion in rats. Other mechanisms might be proposed, such as that hyperglucagonemia could reflect a reduced effectiveness of insulin upon alpha cells as well as upon other sensitive tissues. They also pointed out that hyperaminoacidemia and hyperkalemia might also account for the increased plasma glucagon levels, since these conditions are associated with increased glucagon levels in serum and are also found in diabetic ketoacidosis. On the other hand, Cryer et al. (16) have reported that hypoinsulinemia and hyperglycemia develop promptly during E. coli septicemia in the nondiabetic baboon. They believed that insulin resistance was not a major determinant of the hyperglycemia associated with septicemia. They suggested that the occurrence of hyperglycemia is not an obligatory response to hormones other than insulin. The plasma half-time (a measure of utilization and excretion) of exogenous insulin was significantly prolonged during septicemia in their animals. Thus, they believe the observed fall of endogenous plasma insulin concentrations during E. coli septicemia must represent a de- crease in pancreatic insulin secretion. It is obvious that studies of the metabolic alterations in infected diabetics are incomplete, of great interest, and far from conclusive at present. Response patterns probably vary according to the nature and severity of the infection. This is an important area, and one in which signif- icant advances immediately applicable to the care of diabetics with infection might result. It deserves high priority in future work. Miscellaneous Aspects of Infection in Diabetics Lerner and Weinstein (34) have shown that penicillin is adsorbed abnormally slowly following intramuscular injection in diabetic patients over 50 years of age. Maximum serum levels are lower than in nondiabetic controls, and occur one hour after injection in contrast to one-half hour in control subjects. Total excretion of penicillin in the urine was the same in diabetic and non- diabetic subjects after intramuscular drug administration, but there was a slght delay in oncies tion in the diabetic. Diabetic and normal individuals responded in similar fashion to an intra- venous injection of penicillin, as evidenced by identical serum curves and similar urine excre- tion patterns. The data suggest that delayed adsorption of penicillin from intramuscular depots may be related to diabetic microangiopathy, although other factors have not been excluded. Not all diabetics demonstrated this defect. Similar results were observed with sulfisoxazole, but they were not statistically significant. Under most circumstances this phenomenon probably has no clinical significance. However, it might be important in particularly stubborn or refractory infections. It is therefore noteworthy that the original observations which led to this study concerned two patients with diabetes mellitus and alpha streptococcus endocarditis who failed to respond to ordinary intramuscular doses of penicillin but did respond to greatly increased doses. These pharmacological studies need to be extended to other antibiotics (and indeed to many other important classes of medica- tions). More attention needs to be given to the potential clinical significance of delayed adsorption of drugs in patients with diabetes. Acute Complications 167 The treatment of an infected wound in a diabetic by local application of insulin has been recommended by Paul (47) and others. Many patients with wound infections treated in this way with favorable results have been reported anecdotally, but we are unaware of controlled studies either proving or disproving the worth of this practice. Such a study needs to be done because many physicians, particularly surgeons, utilize this practice empirically. The favorable effect has been assumed to be related to an effect of insulin upon glucose metabolism in the wound, but may be related instead to the presence of zinc in the insulin preparation. Zinc is believed to have profound effects upon wound healing. While a seemingly minor adjunct to treatment of wound infections, the value of this measure could be determined very quickly, and if a definite effect was proven, further interesting studies are suggested. REFERENCES 1. Aguilar-Paradae, Eisentrautam, and RH Unger 1969. Pancreatic glucagon secretion in normal and diabetic subjects. Am J Med Sci 257:415-419. 2. Ainsworth, SK, and F Allison, Jr 1970. Studies on the pathogenesis of acute inflammation. IX. The influence of hyperosmolarity secondary to hyperglycemia upon the acute inflammatory response induced by thermal injury to ear chambers of rabbits. J Clin Invest 49:433-441. 3. Allison, F Jr, and MG Lancaster 1965. Pathogenesis of acute inflammation. VI. Influence of osmolarity and certain metabolic antagonists upon phagocytosis and adhesiveness of leukocytes recovered from man. Proc Soc Exp Biol Med 119:56-61. 4. Bagdade, JD, KL Nielson, and RJ Bulger 1972. Reversible abnormalities in phagocytic function in poorly controlled diabetic patients. Am J Med Sci 263:451-456. 5. Bagdade, JD, K Nielson, R Root, and R Bulger 1970. Host defense in diabetes mellitus: The feckless phagocyte during poor control and ketoacidosis. Diabetes 19:364. 6. Balch, HH, M Watters, and D Kelley 1963. Blood bactericidal studies and serum complement in diabetic patients. J Surg Res 3:199-212. 7. Brayton, RG, PE Stokes, and MS Schwartz, et al. 1970. Effect of alcohol and various diseases on leukocyte mobilization, phagocytosis, and intracellular killing. New Eng J Med 282:123- 128. 8. Briscoe, HF, and F Allison, Jr 1965. Diabetes and host resistance. I. Effect of alloxan diabetes upon the phagocytic and bactericidal efficiency of rat leukocytes for pneumococcus. J Bact 90:1537-1541. 9. Bybee, JD, and DE Rogers 1964. The phagocytic activity of polymorphonuclear leukocytes obtained from patients with diabetes mellitus. J Lab Clin Med 64:1-13. 10. Chernew, I, and AI Braude 1962. Depression of phagocytosis by solutes in concentration found in the kidney and urine. J Clin Invest 41:1945-1953. 11. Cluff, LE, RC Reynolds, DL Page, and JL Breckenridge 1968. Staphylococcal bacteremia and altered host resitance. Annals Intern Med 69:859-873. 12. Cohen, LS, FR Fekety, and LE Cluff 1964. Studies on the epidemiology of staphylococcal in- fection: VI Infections in the surgical patient. Annals Surg 159:321-334. 13. Crosby, B, and F Allison, Jr 1966. Phagocytic and bactericidal capacity of polymorphonuclear leukocytes recovered from venous blood of human beings. Proc Soc Exp Bio Med 123:660-664. 14. Cruickshank, AH 1954. Resistance to infection in the alloxan-diabetic rabbit. J Path Bact 67:323-334. 168 15. 16. 17+ 18. 19. 20. 21. 22. 23. 24, 25. 26. 27- 28. 20. 30. 31. 32. 33. 34. aw Diabetes Mellitus Cruickshank, AH,.and TP Payne 1949. Antipneumococcal powers of the blood in alloxan dia- betes in the rabbit. Bull Johns Hopkins Hosp 84:334-343. Cryer, PE, AG Coron, J Sode, CM Herman, and DL Horwitz 1972. Lethal Escherichia coli septicemia in the baboon: Alpha-Adrenergic inhibition of insulin secretion and its relation- ship to the duration of survival. J Lab Clin Med 79:622-637. DaCosta, JC 1907. The opsonic index in diabetes mellitus. A preliminary record of the find- ings in 22 cases of glycosuria, with remarks on the technique of the opsonin test and on its clinical utility. Am J Med Sci 134:57-70. Drachman, RH, RK Root, and WB Wood, Jr 1966. Studies on the effect of experimental non- ketotic diabetes mellitus on antibacterial defense. I. Demonstration of a defect in phagocytosis. Dubos, RJ 1954. Biochemical determinants of disease. Harvard University Press, p 39. Dumm, ME 1957. Glucose utilization and lactate production by leukocytes of patients with diabetes mellitus. Proc Soc Exp Biol Med 95:571-573. Elder, TD, and RD Baker 1956. Pulmonary mucormycosis in rabbits with alloxan diabetes: Increased invasiveness of fungus during acute toxic phase of diabetes. AMA Arch Path 61:159-168., Esmann, V 1965. Effect of insulin on human leukocytes. Diabetes 12:545-549. Fekety, FR, JR Caldwell, D Gump, JE Johnson, W Maxon, J Mulholland, and R Thoburn 1971. Bacteria, viruses and mycoplasmas in acute pneumonia in adults. Am Rev Resp Dis 104:499-507. Gallin, JI, D Kaye, and WM O'Leary 1969. Serum lipids in infection. New Eng J Med 281: 1081-1086. Gellman, DD, CL Pirani, JF Soothill, RC Muehrcke, and RM Kark 1959. Diabetic Nephropathy: A clinical and pathologic study based on renal biopsies. Med 38:321-367. Hill, RH, JL Dettloff, PG Quie 1973. Leukotaxis, random migration and migration inhibition of leukocytes from patients with juvenile diabetes mellitus. Clin Res 21:306. Hinz, CF, Jr 1956. Properdin levels in infectious and noninfectious disease. Ann N Y Acad Sci 66:268-271. Hutchins, GM, WH Sheldon 1972. The pH of inflammatory exudates in acidotic diabetic rabbits. Proc Soc Exp Biol Med 140:623-627. Jones, CP, B Carter, WL Thomas, RA Ross, and RN Creadick 1947. Mycotic vulvovaginitis and vaginal fungi. A report of 208 patients. Am J Obstet Gynec 54:738-747. Kass, EH 1960. Hormones and host resistance to infection. Bacter Rev 24:177-185. Khurana, RC, D Younger, and JR Ryan 1973. Characteristics of Pneumonia in Diabetes. Clin Res XXI:629, (Abst). Kjellstrand, CM, RL Simmons, FC Goetz, TJ Buselmeier, JR Shideman, B Hartitzsch, and JS Najarian 1973. Renal transplantation in patients with insulin-dependent diabetes. Lancet 2:4-8. Lancaster, MG, and F Allison, Jr 1966. Studies on the pathogenesis of acute inflammation. VII. The influence of osmolarity upon the phagocytic and clumping activity of human leukocytes. Am J Path 49:1185-1200. Lerner, PI, and L Weinstein 1964. Abnormalities of adsorption of benzylpenicillin G and sulfisoxazole in patients with diabetes mellitus. Am J Med Sci 248:37-50. 35. 36. 37. 38. 30. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. ST. 52. 53. 54. 55. 56. Acute Complications 169 Lewis, J, and FR Fekety 1969. Gram-negative bacteremia. Johns Hopkins Med J 124:106-111, Lipscomb, H, HL Dobson, and JA Greene 1959. Infection in the diabetic. Southern Med J 52: 16-23. Loria, RM, and S Kibrick 1973. Coxsackievirus BS infection in the hyperlipemic adult mouse. 13th Inter Conf on Antimicro Agents and Chemo 96 (Abstr), p 19-21, September. Marble, A, HJ White, and AT Fernald 1938. Nature of the lowered resistance to infection in diabetes mellitus. J Clin Invest 17:423. Marco, J, C Calle, D Roman, M Diaz-Fierros, M Illanuae, and I Valverde 1973. Hyperglucagon- ism induced by glucocorticoid treatment in man. New Eng J Med 288:128-131. Martin, SP, GR McKinney, R Green, et al. 1953. The influence of glucose, fructose, and insulin on the metabolism of leukocytes of healthy and diabetic subjects. J Clin Invest 32:1171-1174. Miller, ME, and L Baker 1972. Leukocyte functions in juvenile diabetes mellitus: Humoral and cellular aspects. J Ped 81:979-982. Montgomerie, JZ, GM Kalmanson, and LB Guze 1968. Renal failure and infection. Medicine 47:1-32. Mowat, AG, and J Baum 1971. Chemotaxis of polymorphonuclear leukocytes from patients with diabetes mellitus. New Eng J Med 234:621-626. Muller, WA, GR Faloona, and RH Unger 1973. Hyperglucagonemia in diabetic ketoacidosis. Am J Med 54:52-57. Muller, WA, GR Faloona, and RH Unger 1971. The effect .of experimental insulin deficiency on glucagon secretion. J Clin Invest 50:1992-1999. Oppenheimer, BS, and AM Fishberg 1925. Lipemia and the reticuloendothelial apparatus. Arch Intern Med 36:667-681. Paul, TN 1966. Treatment by local application of insulin of an infected wound in a diabetic. Lancet 2:574-576. . Pek, S, SS Fajans, JC Floyd, Jr, et al. 1972. Plasma levels of glucagon in patients with diabetes mellitus. Diabetes 21:324, Suppl 1. Perillie, PE, JP Nolan, and SC Finch 1962. Studies of the resistance to infection in dia- betes mellitus: Local exudative cellular response. J Lab Clin Med 59:1008-1015. Perla, D, and J Marmorston 1941. Natural resistance and clinical medicine. Boston, Little Brown and Co. ’ Powell, EDU, and RA Field 1966. Studies on salicylates and complement in diabetes. Diabetes 15:730-733. Richardson, R 1933. Immunity in diabetes: Influence of diabetes on the development of anti- bacterial properties in the blood. J Clin Invest 12:1143-1149. Robbins, SL, and AW Tucker, Jr 1944. The cause of death in diabetes: A report of 307 autopsied cases. New Eng J Med 231:865-868. ‘ Sbarra, AJ, W Shirley, and JS Baumstark 1963. Effect of Osmolarity on Phagocytosis. J Bacteriol 85:306-313. Sheldon, WH, and H Bauer 1959. The development of the acute inflammatory response to ex- perimental cutaneous mucormycosis in normal and diabetic rabbits. J Exper Med 110:845-852. Shousha, S, and K Kamel 1972. Nitro blue tetrazolium test in children with kwashiorkor with a comment on the use of latex particles in the test. J Clin Path 25:494-497. 170 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67 . Diabetes Mellitus Smith, JA, JJ O'Connor, and AT Willis 1966. Nasal carriage of staphylococcus aureus and diabetes mellitus. Lancet 2:776-777. Smythe, PM, M Schonland, G Brereton-Stiles, et al. 1971. Thymolymphatic deficiency and depression of cell-mediated immunity in protein-calorie malnutrition. Lancet 2:939-944. Soler, NG, MA Bennett, MG Fitzgerald, and JM Malins 1973. Intensive care in the management of diabetic ketoacidosis. Lancet 1:951-953. Tan, JS, C Watanakunakorn, and JP Phair 1970. Differentiation of defective phagocytosis from impaired intracellular killing by neutrophils with the use of lysostaphin. Presented at 10th Inter Conf on Antimicro Agents and Chemo, p 2, October. Thomas, L 1971. The technology of medicine. New Engl J Med 285:1366-1368. Thornton, GF 1971. Infections and diabetes. Med Clin No Am 55:931-938. Wale, RS, and K Madders 1936. Staphylococcal toxoid in the treatment of diabetes. Brit J Exper Path 17:279-287. Walters, MI, MA Lessler, and TD Stevenson 1971. Oxidative metabolism of leukocytes from nondiabetic and diabetic patients. J Lab Clin Med 78:158-166. Weber, G, NB Stamm, and EA Fisher 1965. Insulin: Inducer of pyruvate kinase. Science 149:65-67. Whitehouse, FW, and HF Root 1956. Necrotizing renal papillitis and diabetes mellitus. J Am Med Assoc 162:444-447. Younger, D 1965. Infections in diabetes. Med Clin No Am 49:1005-1013. Chapter Chapter Chapter Chapter Chapter Chapter 14 15 16 17 18 19 LONG-TERM COMPLICATIONS Demonstrable Metabolic Abnormalities in Diabetes Mellitus that may Contribute to the Pathogenesis of Specific Late Complications A. I. Winegrad and Rex S. Clements, Jr. Diabetes, Hyperglycemia and Atherosclerosis: New Research Directions Leon D. Ostrander, Jr. and Frederick H. Epstein Ocular Complications in Diabetes Morton E. Smith and Bernard Becker Renal Disease in Diabetes Mellitus K. Lundbaek and R. (sterby Diabetic Peripheral Neuropathy J. A. Moorhouse Effect of Diabetes Mellitus on Fetal Growth and Development Daniel H. Mintz and Ronald A. Chez 173 194 213 227 243 256 14 DEMONSTRABLE METABOLIC ABNORMALITIES IN DIABETES MELLITUS THAT MAY CONTRIBUTE TO THE PATHOGENESIS OF SPECIFIC LATE COMPLICATIONS Albert I. Winegrad and Rex S. Clements, Jr. INTRODUCTION The development of effective methods for the prevention of the long-term complications of diabetes mellitus will probably require the identification of the specific biochemical mechanisms responsible for each of the diverse clinical syndromes to which this term is applied, since there is little evidence to support the concept that a single pathological process is responsible for their development. It is recognized that diabetes mellitus may be a group of diseases with heterogeneous etiologies that have as their common characteristics the development of an impaired insulin secretory mechanism and an abnormality in the regulation of plasma glucose concentration that permits diagnosis. Since with rare exception a specific etiology cannot be assigned in any given patient with diabetes mellius, the possibility that specific complications are primarily associated with specific etiologic forms of diabetes remains to be excluded. However, none of the long-term complications of diabetes appears to be restricted to what is presently termed genetically determined diabetes. (Some workers still dispute the association of diabetic retino- pathy and nephropathy with diabetes associated with chronic pancreatitis, hemachromatosis, and pancreatectomy; this is considered in subsequent contributions to this section.) It is therefore reasonable to consider the metabolic abnormalities that are commonly associated with diabetes mellitus to determine whether they may contribute to the development of specific long-term complication. The development of clinical manifestations is clearly preceded by a prolonged period of asymptomatic pathologic change in many of the long-term complications of diabetes mellitus, and the search for detectable metabolic abnormalities that influence their development is an essential part of any program for the development of effective preventive therapy. Until quite recently efforts to determine whether the commonly demonstrable metabolic de- velopment of specific late complications have been restricted to clinical studies whose interpreta- tion remains a matter of dispute. This subject is considered in detail elsewhere in this volume; however, it is appropriate to comment that the data derived from the clinical trials reported to date do not exclude the existence of relationships between the demonstrable metabolic abnormalities in diabetes mellitus and the pathogenesis of any of the late complications. Hyperglycemia Abnormalities in the daily fluctuations in plasma glucose concentration are universally present in patients with clinically detectable diabetes mellitus, and may be present intermittently in some patients over a prolonged period (29). The degree of abnormality in the regulation of plasma “glucose in many diabetics has probably been underestimated since the random blood sugars obtained from these patients are almost invariably evaluated in terms of the blood or plasma glucose con- centrations observed in normal subjects after the ingestion of a large load of pure glucose (usually 50 to 100 grams). Molnar et al. (74) have recently monitored the plasma glucose fluctua- tions that occur over a 24-hour period in normal subjects eating a mixed diet; these fluctuations 173 174 Diabetes Mellitus are restricted to a much narrower range than those observed in normal subjects during an oral glucose tolerance test. Abnormalities in the range and pattern of plasma glucose fluctuations are present in diabetics eating a mixed diet (74); however, these cannot be adequately assessed by isolated plasma glucose determinations or by reference to the values observed in normal subjects during a glucose tolerance test. It should be noted that the value of ''mormalizing" plasma glucose fluctuations in patients with diabetes mellitus has never been assessed in large scale clinical trials and may not be feasible in many patients with the therapeutic methods presently available. Thus the possibility that hyper- glycemia may contribute to the pathogenesis of specific late complications cannot be excluded on the basis of the clinical data presently available. Recent studies suggest that increased activity of the polyol pathway may serve as a model for biochemical mechanisms by which hyperglycemia might produce significant derangements in the metabo- lism of tissues that are the sites of pathological changes in long-term diabetics. The polyol pathway of glucose metabolism consists of two enzymatic reactions by which free (i.e., non- phosphorylated) glucose is converted to free fructose: (1) D-Glucose + NADPH + H'————— 5 Sorbitol + NADP" Alditol:NADP Oxidoreductase (aldose reductase) (2) Sorbitol + NND'———— 5 D-Fructose + NADP + H' L-Iditol:NADP Oxidoreductase (sorbitol dehydrogenase) In this sequence glucose is initially reduced to its polyol (polyhydric alcohol) derivative, sorbitol, through the action of aldose reductase which requires NADPH as a co-factor. Sorbitol is subsequently oxidized to fructose by enzymes resembling hepatic sorbitol dehydrogenase for which NAD® is the preferred co-factor (49, 50). These reactions appear to occur in the cytoplasm, but in most tissues the subcellular localization of the pathway has not been rigorously examined. The polyol pathway is unusual in that glucose utilization by this sequence does not involve the initial formation of glucose-6-phosphate as is the case in the quantitatively major pathways for glucose metabolism in mammalian tissues. Although the individual reactions are potentially reversible, the polyol pathway has been found to operate in an essentially irreversible fashion in those tissues in which it has been most thoroughly studied. This behavior is probably the result of the large decrease in free energy that is associated with cytoplasmic sequences in which NADPH is utilized for the reduction of NAD", which is one of the net effects of glucose utilization via the polyol pathway (2). The polyol pathway provides the mechanism for the synthesis of seminal fluid fructose in the accessory glands of the male genital tract (49, 50). (Fructose appears to be the preferred sub- strate for glycolysis in most mammalian spermatozoa including man.) The polyol pathway also pro- vides the mechanism for the synthesis of fructose in the placenta which is the source of the significant concentrations of free fructose found in fetal plasma (20,50,90). Until quite recently it was believed that aldose reductase, and hence polyol pathway activity, was restricted to these tissues. However, recent studies indicate that aldose reductase activity is widely dis- tributed in mammalian tissues (19) and at present there is evidence that the polyl pathway is normally operative in human seminal vesicles, placenta, lens, brain, peripheral nerve, aortic Demonstrable Metabolic Abnormalities 175 intima and media, and erythrocytes (19,20,27,75,84,90,121). (Other tissues have not been examined as yet.) Although the function of the polyol pathway in tissues other than the seminal vesicles and placenta is obscure, the presence of this pathway in tissues in which the intra- cellular transport of glucose is not subject to insulin regulation and is not rate limiting for glucose concentration may result in increased polyol pathway activity with the development of associated derangements in metabolism and composition. This was first recognized as a consequence of studies of experimental cataract formation by van Heyningen (114). Lens. A number of observations suggest a relationship between elevated plasma levels of glucose and other aldoses (e.g., galactose) and the development of cataracts. ''Snowflake' cata- racts occur primarily in adolescents and infants with symptomatic hyperglycemia (15). Although senile cataracts occur in both nondiabetics and diabetics, cataract extraction is more frequently performed in adult diabetics (i.e., over age 40) (15). Cataracts are also a well recognized complication of human galactosemia either of the classic variety or resulting from galactokinase deficiency (93). In rats with experimental diabetes, cataracts develop with great regularity and there is an inverse relationship between the degree of hyperglycemia and the time required for cataract de- velopment. Lowering the blood sugar by treatment with phloridzin delays or prevents cataract formation without correcting the insulin deficiency (82). Van Heyningen (112, 113, 114) demon- strated the enzymes of the polyol pathway in rat lens, and called attention to the accumulation of the polyol derivatives of glucose, galactose, and D-xylose in the lenses of rats in whom cataract formation resulted from alloxan diabetes, or from feeding diets with high galactose or D-xylose contents. These observations suggested that increased polyol formation is a common feature of experimental ''sugar cataracts. The intracellular transport of glucose is not rate limiting for glucose phosphorylation in the lens, and direct effects of insulin on glucose utilization have not been consistently demon- strated in this tissue (60,113,114). Elevated plasma glucose levels result in increased con- centrations of free glucose in the lens in human and experimental diabetes (84, 113). Lens aldose reductase exhibits a high Km for free glucose and the intracellular glucose concentration is a major determinant of the rate at which it is reduced to sorbitol (47,54,113,114). Although sorbitol is a normal constituent of the lens, its concentration is significantly elevated in both human and experimental diabetes (84,113). A relationship between increased polyol pathway activity in lens and the development of cataracts is well established. Mice of the CFW strain have lenses which are deficient in both aldose reductase and sorbitol dehydrogenase activities; cataracts do not develop when these mice are made diabetic with alloxan despite the presence of high concentrations of free glucose in the lens (59). Similarly the cataracts that develop when rabbit lenses are maintained in tissue culture in the presence of high medium glucose concentrations are prevented by the addition of an inhibitor of lens aldose reductase, 1,l-cyclopentanediacetic acid (18). The association between increased polyol formation and the development of cataracts has been attributed to the osmotic effects of increased intracellular concentrations of sorbitol or other polyol products of aldose reductase activity (54). Tissues other than liver appear to lack facilitated transport systems for sorbitol and other acyclic hexitols, and even in the face of large concentration differences their rate of transport across the cell membranes is very slow 176 Diabetes Mellitus (62, 116). The turnover of sorbitol in the lens has been assumed to be slow, and the magnitude of the increases in lens sorbitol concentration observed in experimental diabetes would exert a significant osmotic effect and account in part for the associated increases in lens water. In- creased medium osmolality minimizes the defects in the active transport of K, amino acids, and myoinositol that have been demonstrated in lenses exposed to high concentrations of galactose under conditions that result in cataract formation; these observations have also been used to support the osmotic hypothesis for the development of sugar cataracts (54, 55, 56). There are, however, significant limitations to the osmotic hypothesis. Recent electron microscopic studies indicate that even in its earliest phases the formation of a sugar cataract involves much more than the simply "hydropic swelling" of the lens fibers which Friedenwald and Rytel (34) observed by light microscopy. The earliest changes occur in the anterior lens epithe- lium which exhibits a marked increase in free and membrane-bound ribosomes and subsequently pro- liferates to form a multicellular layer (61). Similarly, the data of Stewart et al (100) suggest that the increased lens water content in rats fed a high galactose diet may not result solely from the osmotic effects of the accumulation of galactitol in the lens. Erythrocyte. In considering the possible consequences of increased polyol pathway activity in other tissues there has been a tendency to approach the question in terms of the osmotic hypothe- sis. This is scarcely justified on the basis of the data available, since there is evidence that metabolic derangements resulting from increased polyol pathway activity need not be restric- tea to those predicted by the osmotic hypothesis. Thus in human erythrocytes polyol pathway ac- l tivity accounts for approximately 1.8 percent of the glucose uptake at a physiological medium glucose concentration but may increase to 11 percent during incubation with 50 mM glucose (75, 108). Under these conditions, the increased utilization of NADPH for glucose reduction to § sorbitol results in increased pentose phosphate activity as a consequence of changes in the redox | state of the NADP" /NADPH couple. In addition, the markedly increased rate of sorbitol oxidation | to fuctose results in an increased free NADH/NAD* ratio with resultant changes in the steady state ' levels of the glycolytic intermediates proximal to the glyceraldehyde-3-phosphate dehydrogenase re- action and a significant fall in erythrocyte 2,3-diphosphoglycerate (108). Thus in the human erythrocyte increased polyol pathway activity may significantly alter the redox state of both free {' pyridine nucleotide couples. It is of interest that van Heyningen originally suggested that such changes might contribute to the development of experimental ''sugar' cataracts (113). Aortic Wall. The enzymes of the polyol pathway have been demonstrated in human thoracic aorta and in the intima and media of rabbit thoracic aorta (19, 76). The latter tissue provides a convenient in vitro system for the study of the effects of elevated glucose concentrations on {} the metabolism of the arterial wall (118). Intracellular transport does not appear to be rate i || limiting for glucose utilization in the arterial wall since free intracellular glucose can be rn in tissue from both normal and alloxan diabetic rabbits that have been incubated PETE { with a physiological glucose concentration (5mM) (118). In addition, the free intracellular “glucose concentration has been shown to rise following incubation with elevated glucose concentra- tions (76). Direct or immediate effects of insulin on glucose metabolism have not been demon- strated in aortic intima and media with the exception of minimal increases in the recovery of 14%¢ from 14C-glucose in glycogen (118). Increasing the medium glucose concentration from 5 mM to 20 mM, which is within the range observed in the plasma of human diabetics, results in a Demonstrable Metabolic Abnormalities 177 threefold increase in polyol pathway activity in the aorta (76). Polyol pathway activity is best assessed by the determination of aortic fructose production, since in this tissue as well as in the human erythrocyte there is a very significant turnover of endogenously synthesized sorbitol (75, 76). = Increased polyol pathway activity results in striking alterations in the composition and metabolism of aortic intima and media. Incubation with 20 mM or higher glucose concentrations | results in an increase in the water content of the aorta within two hours that is similar in if magnitude to that observed in the aortae of chronically hypertensive rats (76, 107). This occurs in the face of a decrease in the extracellular volume of the aorta as measured by the inulin space. The increase in aortic water content is too large to be explained as a consequence of the observed increases in aortic sorbitol content (76). However, there is no doubt that it is a con- sequence of increased polyol pathway activity since it is prevented by the inhibition of aortic aldose reductase and polyol pathway activity with 1,1-cyclopentanediacetic acid (77). Moreover, incubation with elevated concentrations of galactose, which is also a substrate for aortic aldose reductase, also results in increases in aortic water content (76). Although the aorta characteristically exhibits a high rate of aerobic glycolysis (i.e., glucose conversion to lactate), it has a significant oxygen uptake and a modest Pasteur effect i can be demonstrated (115). Acute exposure to elevated glucose concentrations does not alter the oxygen uptake of aortic intima and media, but within 30 minutes, a period sufficient to result in a detectable increase in water content, the oxygen uptake begins to decline (76). After a two-hour | pre-incubation with 20 mM or higher glucose concentrations, the oxygen uptake of the aorta is , significantly less than that of paired samples incubated with 5 mM glucose, its rate of lactate J} production is increased, and there is an increase in aortic lactate/pyruvate ratio reflecting a 1 change in the free NADH/NAD* ratio. These changes, which suggest inpaired oxygenation of the ig ! l-aortic wall, occur in medium saturated with a physiological oxygen tension. Increasing the oxygen “tension (by saturating the medium with 95 percent oxygen) restores oxygen uptake and lactate pro- duction to levels similar to that observed in tissues incubated with 5 mM glucose (76). Thus exposure to elevated glucose concentrations appears to result in hypoxia of aortic intima and media which appears to result from impaired oxygen diffusion. The addition of 1,1-cyclopentandia- cetic acid in concentrations sufficient to produce a 40 percent inhibition of polyol pathway activity in aortic tissue incubated with 20 mM glucose prevents the development of both the in- crease in aortic water content and the associated impairment in oxygen diffusion at a physio- logical oxygen tension (77). These in vitro effects of elevated glucose concentrations on the composition and metabolism of the aortic wall may have in vivo counterparts since both an increase in water content and impaired respiration have been demonstrated in the aortic intima and media of alloxan diabetic rabbits (76). The nature of the alterations in aortic metabolism that result from exposure to elevated glucose concentrations and increased polyol pathway activity is of interest in light of current speculation concerning factors that may adversely affect the response of this tissue to exogenous factors operative in the pathogenesis of arterial lesions. It has been noted that the arterial intima and media of larger vessels are devoid of capillaries and are dependent upon diffusion from the lumen for the provision of their oxygen requirements (115). Further, it has been noted that the ability to provide the oxygen requirements of aortic intima and media by diffusion 178 Diabetes Mellitus is limited by the distance through which this process must occur, and that minimal increases in diffusion distance would be expected to result in hypoxia of the aortic wall in the face of a physiological oxygen tension in arterial blood (115). Haust (45), Robertson (88), and Whereat (115) have all speculated on the possible significance of hypoxia of the arterial wall in po- tentiating the effects of other factors operative in inducing arterial lesions. Whether hypoxia is a significant factor in the proliferation of arterial smooth muscle, which Ross and Glomset (91) have postulated to be the key event in the pathogenesis of atherosclerotic lesions remains to be determined. It is, however, apparent that hyperglycemia itself can induce significant alterations in the metabolismand composition of the arterial wall that are similar to those which are thought to favor the development of arterial disease. Peripheral Nerve. Current evidence suggests that the development of the symmetrical distal polyneuropathy associated with diabetes mellitus results from a metabolic derangement rather than from occlusive disease of nutrient blood vessels. By sufficiently sensitive techniques widespread abnormalities in peripheral motor and sensory nervous function can be demonstrated in newly diagnosed diabetics without evidence of clinical neuropathy (17). These abnormalities in- clude decreased motor and sensory nerve conduction velocity. The latter are usually interpreted as indicating a derangement in myelin function, i.e., segmental demyelination or changes in the resistance or capacitance of internodal and paranodal myelin. Insulin deficiency or hyperglycemia may contribute to the changes that occur in peripheral nerve for experimental diabetes produced by pancreatectomy, alloxan, or streptozotocin results in reduced conduction velocity in rat sciatic nerve and in changes in the electrical properties of its myelin sheath (25,35). The polyol pathway is operative in peripheral nerve and elevated levels of glucose, sorbitol and fructose are present in the peripheral nerves of rats with experimental diabetes (37,101,103, 104. Polyol pathway activity in peripheral nerve appears to be regulated in part by ambient glucose concentration, and lowering the blood sugar of alloxan diabetics by the administration of insulin results in a rapid fall in the glucose, sorbitol and fructose contents of peripheral nerve (104). The localization of polyol pathway activity within the peripheral nerve has not been clearly established; however, the sharp rise in the fructose content of peripheral nerves under- going Wallerian degeneration at the time which coincides with rapid multiplication of Schwann cells suggests that a significant fraction of polyol pathway activity is localized within those cells (102). The intrinsic relationship between the Schwann cell and the myelin sheath of myelinated nerves has led to speculation that derangements in the metabolism of the Schwann cell resulting from increased polyol pathway activity may contribute to the pathogenesis of diabetic neuropathy (36.37)... As yet there has been no clear demonstration that hyperglycemia as distinct from insulin deficiency can induce abnormalities in peripheral nervous function. Plasma Lipids and Lipoproteins It is commonly assumed that atherosclerotic disease in diabetics is accelerated as a conse- quence of associated hyperlipidemia (see Section on Atherosclerosis). However, the evaluation of plasma lipids in diabetics is difficult because of the existence of independently determined fac- tors that may contribute to any observed abnormality. This point has been clearly established by the recent studies of Goldstein et al. (41, 42, 43) on the genetics of the hyperlipidemia observed in survivors of myocardial infarction. Hyperlipidemia was present in 33 percent of 500 survivors Demonstrable Metabolic Abnormalities 179 of a myocardial infarction. By detailed family studies three distinct inherited lipid disorders (familial hypercholesterolemia, familial hypertriglyceridemia, and a newly recognized entity, familial combined hyperlipidemia) were shown to be present in 20 percent of the survivors less than 60 years of age, and in 7 percent of all older survivors. These studies clearly indicate that a high fraction of the hyperlipidemia associated with coronary artery disease is genetically determined. The frequency of diabetes mellitus in the 500 survivors was examined using a fasting blood sugar greater than 120 mg percent or current therapy with insulin or an oral antihyper- glycemic agent as the diagnostic criteria. By these rigorous criteria 12 percent of the sur- vivors were found to be diabetic. The frequency of diabetes in these survivors in whom an in- herited lipid disorder had been established by family study was 22 percent in those with familial hypertriglyceridemia, 15 percent in those with combined hyperlipidemia, and 6 percent in those with familial hypercholesterolemia. Although the criteria employed would probably have under- estimated the true frequency of diabetes in these survivors, it would appear that diabetes mellitus frequently co-exists with familial hypertriglyceridemia in survivors of myocardial in- farction and that the frequency with which it co-exists with familial combined hyperlipidemia in such populations deserves further study. Family studies are currently required to document the presence of a genetically determined disorder of lipid metabolism since the lipoprotein pheno- types (32) do not appear to be qualitative markers in a genetic sense but rather quantitative parameters which may vary among different individuals with the same genetic lipid disorder (48). The lack of a specific genetic marker for diabetes mellitus and the failure to define the distri- bution of genetically determined lipid disorders in the patients studied seriously restricts the interpretation of most of the data presently available concerning plasma lipids and lipoproteins in diabetic populations. Most diabetics have normal serum triglyceride and cholesterol concentrations (78,99,120). The highest reported incidence of hyperlipidemia (either hypercholesterolemia or hypertrigly- ceridemia) in diabetics approximates 30-35 percent when compared with arbitrary limits derived from age and sex matched normals (9). The fact that diabetes mellitus and hyperlipidemia co-exist more frequently than would be expected by chance is generally accepted. However, although some have assumed that a causal relationship must exist, this has thus far been demonstrated only in restricted instances. In ketoacidosis the serum triglycerides and cholesterol levels may be markedly elevated and in these instances there is a marked increase in the plasma very low denisty lipoprotein (VLDL) concentration which may be accompanied by increases in chylomicra (24,46). These abnormalities in plasma lipids are corrected by treatment with insulin. Hyperlipidemia resembling that observed in diabetic ketoacidosis can be induced in insulin-dependent juvenile diabetics by the cessation of insulin administration (5). The hypertriglyceridemia associated with ketoacidosis has been attributed, in part, to an acquired deficiency of lipoprotein lipase activity (LPL). VLDL and chylomicra are triglyceride-rich lipid particles which appear to share a common saturable removal mechanism (13). The hydrolysis of the triglycerides of chylomicra and VLDL is mediated by LPL, which is found in particularly high concentrations in the capillary endothelium of adipose tissue (31). In rats experimental diabetes results in a marked reduction in adipose tissue LPL activity, which is corrected by treatment with insulin (51,89). Tissue LPL activity is difficult to assess in intact man. The intravenous administration of heparin appears to displace LPL from 180 Diabetes Mellitus tissue sites to which it is bound, and post-heparin plasma catalyzes the hydrolysis of tri- glycerides in chylomicra and VLDL. This post-heparin lipolytic activity (PHLA) acts on artificial triglyceride emulsions and assays of this activity have been used as an indirect measure of LPL (33). Insulin deficiency has also been implicated in the marked increases in plasma triglyceride- rich particles, primarily chylomicra, that are infrequently observed in chronic symptomatic diabetes with minimal ketosis (3, 4). These patients exhibit subnormal PHLA which can be restored to normal by treatment with adequate quantities of insulin; this is accompanied by improved tri- glyceride removal and a reduction in plasma triglyceride levels. In less severe diabetics with mild fasting hyperglycemia PHLA is usually normal; however, in some of these individuals with associated hypertriglyceridemia the LPL-related maximal tri- glyceride removal rate (as measured in vivo during prolonged heparin infusion) is low (9,11,85). Brunzell et al. (13, 14) found that diabetics with normal PHLA but with impaired triglyceride removal exhibit a fall-off in PHLA activity after the second hour of a 5-hour infusion of heparin, which differs from the normal sustained response to heparin infusion. They have related this "PHLA depletion" to varying degrees of insulin deficiency, but as yet the reversibility of this abnormal response to heparin infusion following insulin treatment has not been fully documented. Thus in ketoacidosis and in the two instances described above, it is clear that insulin deficiency can contribute directly to the development of hypertriglyceridemia. Hypertriglyceridemia is also observed in diabetics without fasting hyperglycemia and often with only mild impairment of glucose tolerance. Although the increase in plasma triglyceride levels which results from the substitution of a high carbohydrate diet for a normal diet appears to be exaggerated in this group, the response is now considered neither unique nor abnormal (9). In some of these subjects an abnormally high triglyceride influx rate into the circulation can be demonstrated (11,79,86). No abnormalities of PHLA or triglyceride removal kinetics have been uncovered, but fractional removal rates may be reduced (9). Although both free fatty acids and glucose are major precursors for endogenous triglyceride production, there is no evidence that substrate flux is accelerated in the type of individuals under consideration. The lipoprotein pattern most frequently associated is one in which pre-B-lipoproteins are the only triglyceride- rich lipoproteins present in increased amounts, however, in patients with the highest pre-B- lipoprotein concentrations chylomicronemia may also be present. This is believed to result from a saturation of the LPL-related plasma triglyceride removal system (9,13,14,86). The causes of the hypertriglyceridemia associated with mild glucose intolerance are unknown. Since the necessity for family studies to establish the existence of the common inherited dis- orders associated with coronary artery disease has only recently been recognized most of the available data is derived from studies of populations that are not sufficiently well characterized to permit firm conclusions. It has been suggested that the association between hypertri- glyceridemia and mild glucose intolerance may be explained by the common coexistence of obesity (10, 12). Correlations between plasma triglyceride concentrations and obesity have been observed frequently (1,28,30,44,52,87). Obesity is known to be associated with true hypersecretion of insulin in both normal and diabetic subjects, i.e., the insulin secretory response to com- parable stimuli is greater than that seen in normal weight nondiabetic and diabetic subjects (57). In a variety of populations serum insulin, both basal and after stimulation, correlates with plasma triglyceride concentrations (12,86,92). (Although lipoprotein phenotypes are nonspecific, Demonstrable Metabolic Abnormalities 181 Glueck et al. (40) failed to observe a positive correlation between insulinogenic indices and serum triglycerides in 80 patients with Type III, IV, or V patterns; obesity was not a signifi- cant factor in these patients but a high percentage had abnormal glucose tolerance.) This has led to speculation that insulin may be one of several factors promoting hepatic triglyceride sythesis (68) and thus aggravating the accumulation of triglyceride-rich lipoproteins in the plasma of subjects with endogenous lipemia (9,12,79). Weight reduction has been shown to result in correction of glucose intolerance, '"hyperinsulinism," and hypertriglyceridemia in some individuals (12). However, defective early insulin responses to glucose are similar in hyper- triglyceridemic and normolipidemic subjects with comparable degrees of mild glucose intolerance (6); this has led to the suggestion that early diabetes with its associated abnormalities in early insulin release is not etiologically related to hypertriglyceridemia (9). It is apparent that the classification and delineation of the abnormalities in lipid metabolism present in the diabetic population is still in a developmental stage. However, the frequency with which inherited abnormalities in lipid metabolism are encountered in survivors of myocardial infarction, and the frequency with which some of these disorders coexist with diabetes mellitus (as defined by fairly rigorous criteria) suggests that failure to define the distribu- tion of such patients seriously limits the value of all the reported clinical trials in so far as cardiovascular mortality is concerned. It should be noted that Havel (46) has expressed doubt that the increased morbidity and mortality from ischemic coronary and peripheral vascular disease in diabetic subjects can be accounted for solely by an augmented rate of atherogenesis (see Section on Atherosclerosis). Glycoproteins and Mucopolysaccharides Winzler (119) has reviewed the extensive literature which indicates that the level of total serum protein-bound carbohydrate tends to be elevated in human diabetics. Most of the studies which attempted to relate alterations in serum glycoproteins and the presence of vascular complica- tions in diabetics are difficult to interpret; this results from the large number of glyco- proteins in plasma and the failure to isolate and quantify specific glycoprotein components (119). Thus, although elevation of the serum ojp-globulin fraction has been observed in diabetics, the com- ponents of this fraction may vary independently, and a correlation between an elevated oj-globulin fraction and diabetic retinopathy or nephropathy has not been clearly established (69,119). Ele- vated concentrations of serum o-globulin glycoprotein components have been noted by Cleve et al. (21). McMillan (70) has recently reported a qualitative alteration in the composition of aj;-acid glycoprotein in diabetic serum; this purified component was reported to have an increased content of fucose. The data on plasma mucopolysaccharides in diabetics are fragmentary. Earlier speculation that the lesions in the retina or renal glomeruli of diabetics might result from the deposition of plasma glycoproteins has been discounted (119). The present evidence on the composition of the glomerular basement membrane in long-standing diabetics indi- cates that it clearly differs from that of circulating glycoproteins (7, 8, 95). The liver is known to be the source of most of the plasma glycoproteins with the exception of the immunoglobulins and the glycoprotein hormones (95). Insulin deficiency does not sig- nificantly alter the utilization of glucose for the synthesis of the glucosamine components of glycoproteins in rat liver even though glucose utilization for glycogen synthesis is markedly impaired (94). This and related observations led Spiro (94) to suggest that glucose utilization 182 Diabetes Mellitus for the synthesis of the carbohydrate components of glycoproteins (most of which are derived from glucose) might be increased in human and experimental diabetes as a consequence of decreased glucose utilization by insulin-dependent pathways and the increased availability of glucose. Recent evidence suggests that the glomerular basement membrane of long-standing human diabetics contains increased quantities of a disaccaride containing galactose and glucose linked to the hydroxyl group of hydroxylysine residues (78). The biochemical mechanism(s) responsible for these changes have not been clarified; however, Spiro and Spiro (96, 97, 98) have presented evi- dence that the activity of the ''glucosyltransferase'" of rat kidney cortex which catalyzes the transfer of glucose from uridine diphosphoglucose to galactose linked to hydroxylysine is in- creased in alloxan diabetic rats when compared with age-matched controls, and can be reduced by treating the diabetic rats with insulin. (The structure and synthesis of glomerular basement membrane are considered in detail in a subsequent section of this volume.) Earlier speculation that insulin deficiency and/or hyperglycemia might favor the addition of carbohydrate units to specific amino acids in peptides or proteins has found unexpected support in recent studies of hemoglobin A The mean values for hemaglobin A c are nearly twofold greater in diabetics than in er subjects (109). Structurally odoin Ads a condensation product between one molecule of hemoglobin A (the major component of hemoglobin in normal adult erythrocytes) and one or more hexoses. The presence of increased concentrations of hemoglobin Ale does not appear to be related to the patient's age, duration of disease, or the presence of specific clinical complications (109). However, in contrast to hemoglobin A, the oxygen affinity of hemoglobin A E does not appear to be significantly influenced by 2,3-diphosphoglycerate 1 which is an important regulator of the oxygen affinity of normal erythrocytes. Present evidence indicates that the increased concentrations of hemoglobin A . found in some diabetics represents 1 a modification of normal hemoglobin as a consequence of the diabetic state. Dixon (23) has recently concluded that the glucosylvalylhistidine (Glc-Val-His) formed by the reaction of glucose and Val-His (in pyridine-acetic acid at pH 6.2) corresponds to the substance found to be released from the amino terminus of the o-chains of hemoglobin A 2 by Schroeder. Dixon (23) has suggested that the dissociation constant of this reaction is - under the conditions existing in human erythrocytes no more powerful glycosylating agent than an increased blood glucose level need be postulated to explain the increased concentrations of hemoglobin Ale in diabetics. These studies also suggested that a low degree of glycosylation might possibly be expected for other a-amino groups that are exposed to an elevated glucose concentration (23). Since the glycosyla- tion of hemoglobin A to A ™ has been shown to be associated with a loss of normal regulatory func- 1 tion, the possibility that similar alterations in the structure and function of other biologically important peptides in the diabetic obviously requires consideration. However, no systematic study of this problem has, as yet, been undertaken. There are a number of obvious candidates for study including the apoproteins of the circulating lipoproteins. Hormones There has been recurrent speculation that abnormalities in growth hormone secretion may con- tribute to the pathogenesis of diabetes mellitus or its complications; however, the early studies of plasma human growth hormone concentrations (HGH) by radioimmunoassay (122) failed to demon- strate any significant abnormality in fasting HGH or in the response to glucose ingesion in diabetics. (Clinical trials of the value of hypophysectomy in the management of diabetic Demonstrable Metabolic Abnormalities 183 retinopathy are considered in another contribution to this volume; these studies provide no direct evidence that an abnormality in growth hormone secretion is a significant factor in the pathogenesis of this condition.) Lundbaek (63, 64, 65) has proposed that an abnormality in growth hormone secretion is a causal factor in the development of ''diabetic angiopathy.' These studies have been primarily concerned with documenting the existence of an abnormality in HGH secretion in human diabetics. Plasma HGH in normal adult subjects may exhibit wide acute fluctuations (16,39). Known stimuli to growth hormone secretion include fasting hypoglycemia, a rapidly falling blood glucose concentration in the absence of hypoglycemia, physical exercise, surgical and emotional stress, the ingestion of a protein meal, and the intravenous infusion of specific amino acids (arginine is particularly effective). In addition, a significant peak of plasma HGH often occurs at the onset of deep sleep (16,22,39). Lundbaek and his co-workers (63, 64) compared the HGH levels in young newly diagnosed male diabetics and in normal male medical students in serum samples obtained at 30-minute intervals through a 24-hour period in which a standardized pattern of eating, physical exercise, smoking, and bed rest was imposed. The average serum HGH in the diabetics was 3 to 4 times higher than that of the controls, and the pattern of serum HGH fluctuations in the diabetics was characterized by the presence of many more peaks throughout the day, with little apparent relationship to fluctuations in blood glucose. The same workers reported that a controlled exercise stress which failed to induce a significant rise in serum HGH in normal male controls caused a rise in male diabetics with disease of recent onset to 30 years' duration. Following treatment, which lowered the fasting blood glucose levels to a normal range (60 to 100 mg percent), newly diagnosed male juvenile diabetics exhibited a decreased HGH response to standardized physical exercise (63, 64). Knopf et al. (58) have reported that there was a modest positive correlation between fasting blood glucose and plasma HGH concentrations in 315 patients who attended a diabetic outpatient clinic over a six-month period. The fasting plasma HGH in diabetic males was significantly higher than that of young male controls (58). Female diabetics with retinopathy had a mean fasting HGH that was significantly higher than that of normal young females, and higher, although not significantly so, than that of female diabetics who did not have retinopathy. In male diabetics the presence of retinopathy was associated with a significantly higher fasting level of both plasma HGH and blood glucose (58). Although the normal response to an oral glucose load is an initial fall in plasma HGH, a rise has been observed in association with a variety of diseases as well as in the normal newborn in- fant (32). A number of investigators had suggested that a similar response to glucose ingestion is observed in mildly diabetic patients (123) and in juvenile diabetics (64). Knopf et al. (58) found that this ''paradoxical' response of plasma HGH to glucose ingestion is observed in 10.5 per- cent of 57 healthy controls, in 8.3 percent of 84 patients with latent diabetes, and in 14.8 percent of 27 subclinical diabetics (positive family history of the disease, normal glucose tolerance * but abnormal cortisone glucose tolerance). The 'paradoxical' increase was not a reproducible response in any given patient and was not observed following the intravenous administration of glucose to normal control or latent diabetic subjects. Knopf (58) concluded that the pattern of suppression of HGH secretion during hyperglycemia induced by the administration of glucose is similar in patients with latent diabetes, in close relatives of diabetics, and in healthy controls. 184 Diabetes Mellitus The existence of a consistent abnormality in plasma HGH fluctuations in diabetics requires further documentation. The data available, however, do suggest that an elevated plasma HGH is more frequently present in male diabetics and in female diabetics with retinopathy than in young healthy subjects of the same sex. Further, it would appear unlikely that a fixed abnormality in the regulation of HGH secretion is present in human diabetics. There is some evidence that the degree of abnormality may be related to the range of blood glucose fluctuations present in that patient and can be altered by a closer approximation of the range of blood glucose fluctuations present in normal subjects. The causes for the apparent association between increased fasting plasma HGH levels and the presence of retinopathy remain to be determined, but a causal relation- ship has not been suggested (58). Merimee et al. (72, 73) have noted the frequent occurrence of abnormal glucose tolerance in sexual ateliotic dwarfs who have a monotrophic deficiency of human growth hormone. These workers distinguished two groups: (Type I) exhibited a decreased plasma insulin (IRI) response to the ingestion of glucose, a mixed glucose-beef meal, or to the intravenous infusion of L-arginine; (Type II) had a normal or greater than normal increase in plasma insulin after these stimuli. (The question of what would constitute appropriate controls for these responses has been fre- quently raised.) Since the insulin responses to provocative stimuli and glucose tolerance im- proved in Type I dwarfs given HGH replacement therapy, Merimee (71) concludes that these dwarfs do not have genetic diabetes. Retinopathy has not been observed in sexual ateliotic dwarfs (71). From collaborative studies with Siperstein (73) it was concluded that sexual ateliotic dwarfs do not exhibit abnormally thickened muscle capillary basement membranes. Merimee (71) has concluded that carbohydrate intolerance and the other metabolic abnormalities associated with diabetes are relatively unimportant in the development of microangiopathic lesions in the absence of genetic diabetes and the chronic absence of HGH. The unique populations studied by Merimee are of great interest and deserve further exploration. His conclusions, however, have not gone undisputed. Williamson (117) has noted the potential fallacy of equating microaneurysms in the retinal circula- tion and capillary basement membrane thickening, since the processes may be distinct and subject to independent degrees of modification by growth hormone deficiency. Further, he has suggested that the values for muscle capillary basement membrane in the ateliotic dwarfs are suggestive of a bimodal distribution, and that with one exception the dwarfs with the thickest CBM had the most abnormal glucose tolerance tests. He concludes that it is not justified to conclude that an associa- tion between hyperglycemia and increased capillary basement membrane thickness may not exist in these dwarfs (117). Although abnormal glucose tolerance is observed in a significant number of patients with acromegaly, it has not been possible to draw meaningful conclusions concerning the possible contribution of excess HGH secretion to the pathogenesis of complications in diabetes from the data available. Hyperinsulinism. The existence of true hyperinsulinism in association with diabetes mellitus is disputed (57). As noted in the discussion of lipid abnormalities in diabetics, the relation- ship between diabetes mellitus, obesity, hyperinsulinism and hypertriglyceridemia requires clarification. The situation is further complicated by the suggestion (80, 81) that one can identify within the nonobese survivors of myocardial infarction a group of individuals with "true hyperinsulinism' without impaired glucose tolerance. Similar observations have been reported by Tzagournis et al. (110, 111), Peters and Hales (83), and Gertler et al. (38). As noted by Demonstrable Metabolic Abnormalities 185 by Nikkili et al. (81) all evidence for the possible role of hyperinsulinemia in the patho- genesis of arterial disease is thus far derived from cross-sectional studies and is more sug- gestive than conclusive. However, Stout (105, 106) has presented evidence which he interprets as indicating that insulin stimulates the synthesis of lipids in the arterial wall; the significance of these observations is disputed (66). A lipase has been identified in arterial tissue which appears to be distinct fromlipoprotein lipase and whose activity with tributyrin or tripalmityn as substrate is increased by 3',5'-cyclic AMP, adrenaline, glucagon, cortisol, and growth hor- mone (66, 67). In rat aorta this activity is markedly increased following the induction of alloxan diabetes and is reduced by pretreatment of the animals with insulin (66). In human arterial tissue insulin decreases the lipolytic activity observed in homogenates of aorta, femoral, and coronary intima and media exposed to 5 x 10 > M norepinephrine (67). It has been suggested that hyperinsulinemia during the late postprandial phase may impair the disposition of lipids de- rived from the plasma in the arterial wall (66, 67). Other Hormones. Abnormalities in the secretion of other hormones in diabetes mellitus are considered in other sections of this volume. Data relating such abnormalities to the pathogenesis of the late complications of diabetes are at present fragmentary. Information Required through Research Now and in the Future A. An essential deficiency which must be corrected is the lack of information concerning the normal metabolism of the tissues that are the sites of pathological lesions underlying the late complications of diabetes and the manner in which they are affected by human and experimental diabetes. The studies required include: (1) Delineation of the normal metabolism of human and mammalian capillaries including those found in the retinal, renal glomerular, and peripheral circulations. (2) Delineation of the metabolism and normal function of the components of the capillary including the pericyte (termed mural cells in the retina). (3) Delineation of the processes of capillary basement membrane synthesis, modification, and degradation in the renal glomerular, retinal, and peripheral circulations. (4) Delineation of the normal metabolism and function of the components of the arterial wall: the factors regulating myoepithelial cell migration and proliferation; the synthesis and degrada- tion of elastin, glycosaminoglycans, and collagen within the arterial wall; the means by which lipoproteins cross the endothelial barrier and their subsequent disposition in the arterial wall. (5) Delineation of the normal metabolism of the Schwann cell and axon and their metabolic interrelationships. (6) Direct examination of the manner in which the above are affected by insulin and hyper- glycemia, and in tissues obtained from human diabetics and animals with various forms of genetically determined and pharmacologically induced diabetes. B. Information is required concerning the distribution within the whole population of pa- tients with diabetes mellitus of independent, genetically determined factors that may modify the course with regard to the development of specific complications. (1) Delineation of the frequency with which the recognized inherited derangements in lipid metabolism associated with coronary heart disease coexist with diabetes mellitus. Delineation of the frequency of clinical events related to atherosclerotic vascular disease in diabetics with a 186 Diabetes Mellitus coexistent inherited abnormality in lipid metabolism as compared with nondiabetics with the same inherited defect, and diabetics in whom the presence of an inherited lipid disorder has been ex- cluded as best as possible. (2) Efforts to delineate other subgroups of diabetics with presently unrecognized coexistent independent factors influencing the course of diabetes mellitus. “4 a. Studies to determine whether the development of rapidly progressive diabetic retinopathy, nephropathy, or neuropathy is more frequently associated with diabetes of presumed infectious etiology or with other diseases; efforts to determine whether these events are more frequently encountered within specific family groups. b. Continued pursuit of leads derived from the studies outlined in A to determine whether the development of specific late complications may be determined by coexistent independent factors. To cite specific examples: When the processes of basement membrane synthesis and degradation are better understood, it should be possible to examine the possibility that inherited variations in the enzymes concerned with basement membrane synthesis or degradation exist in humans; the extent to which these might account for the development of specific microvascular complications in the diabetic population could then be examined. Similarly in the future it should be possible to determine whether in- herited variation in polyol pathway activity within specific tissues may decrease or increase the likelihood of the development of specific complications. c. Information is required concerning the normal range of plasma glucose fluctuations throughout the day in nondiabetic males and females of varying ages eating normal mixed diets, as well as the alterations that are associated with obesity, subclinical, and latent diabetes. d. The relationships between diabetes mellitus, abnormalities in plasma lipids, obesity, diet, and altered patterns of insulin secretion must be clarified. e. A continued search is required for evidence of possible relationships between hormonal abnormalities in the diabetic and the development of specific complications. Personnel and Financial Requirements (1) It is apparent that much of the basic information required has immediate application to major health problems other than diabetes mellitus and may to some extent overlap the require- ments for cardiovascular disease, blindness, renal disease, and peripheral nervous disorders. However, traditionally the study of the metabolism of individual tissues has developed in large part as a consequence of studies concerned with diabetes, and this combined interest is frequently encountered in well qualified personnel. (2) Much of the information required is in essence applied biochemistry, and few major bio- chemistry departments would presently consider the metabolism of isolated tissues and their alterations in human and experimental diabetes appropriate research activities for a significant fraction of their senior faculty. It is therefore necessary to continue to develop individuals with suitable training who will enter this field and to provide means to maintain their continued work in this area; provisions are required both for individuals who will restrict their studies to laboratory investigations and need not have clinical training, and for physician investigators. Value of the Proposed Research The diabetic population appears to be heterogeneous not only with regard to etiology but also with regard to the existence of independent factors that may alter the probability that a specific late complication of diabetes will develop in time. It would therefore seem unlikely that any Demonstrable Metabolic Abnormalities 187 single therapeutic approach will prove effective for all diabetics. The proposed research should permit an understanding of the processes responsible for the development of specific late complica- tions and the identification of possible contributing factors. This should permit a segregation of the diabetic population into well defined groups. This is essential for meaningful clinical trials. Moreover, the proposed research should provide an indication of the specific biochemical processes which must be influenced if the development of complications is to be prevented, and will provide systems in which a search for effective pharmacological agents can be carried out. The extent to which the metabolic consequences of an impaired insulin secretory mechanism, including hyperglycemia, may contribute to the development of the late complications can only be assessed in studies of well defined populations. Current efforts to develop an artificial pancreas and to explore the feasibility of islet transplantation must be justified, in part, on the assump- tion that a correction of the metabolic abnormalities associated with diabetes mellitus will in- fluence the development of specific complications. The development and testing of these modali- ties of therapy should be accompanied by efforts to define the appropriate populations in which such therapy may be warranted, and in which the value of these therapies can be appropriately tested. Information on the normal range and pattern of plasma glucose fluctuations in normal life is an obvious prerequisite for such studies. The significance and origin of lipid abnormalities in diabetics in whom there is not a coexisting genetic disorder in lipid metabolism must be defined before appropriate studies to determine the need for specific forms of therapy and to test the value of any therapy. The large number of patients with diabetes mellitus, the chronic nature of the disease, and the time required for the development of related late complications suggest that any economically feasible program to prevent complications will require the development of reliable methods of predicting the likelihood that specific complications will develop. In this manner emphasis can be placed on those subgroups of diabetics in whom the risk is greatest. An essential pre- requisite to the development of these methods is the data to be derived from the proposed studies. REFERENCES 1. Albrink, MJ, and JW Meigs 1967. Interrelationship between skinfold thickness, serum lipids, and blood sugar in normal men. Amer J Clin Nutr 20:777. 2. Atkinson, DE 1971. Metabolic Pathways, 3rd ed., edited by HJ Vogel. New York, Academic Press, vol V, p 1. 3. Bagdade, JD, D Porte, Jr , and EL Bierman 1967a. Diabetic lipemia, a form of fat-induced lipemia. New Eng J Med 276:427. 4. Bagdade, JD, EL Bierman, and D Porte, Jr 1967b. The significance of basal insulin levels in the evaluation of the insulin response to glucose in diabetic and non-diabetic subjects. J Clin Invest 46:1549. 5. Bagdade, JD, D Porte, Jr , and EL Bierman 1968. Actue insulin withdrawal and the regulation of plasma triglyceride removal in diabetic subjects. Diabetes 17:127. 6. Bagdade, JD, EL Bierman, and D Porte, Jr 1971. The influence of obesity on the relationship between insulin and triglyceride levels in endogenous hypertriglyceridemia. Diabetes 20:664. 7. Beisswenger, PJ, and RG Spiro 1970. Human glomerular basement membrane: Alterations in diabetes mellitus. Science 168:596. 188 Diabetes Mellitus 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. Beisswenger, PJ, and RG Spiro 1973. Studies on the human glomerular basement membrane: Composition, nature of the carbohydrate units and chemical changes in diabetes mellitus. Diabetes 22:180. Bierman, EL 1973. Vascular and neurological changes in early diabetes, edited by RA Camerini-Davalos, and HS Cole. New York, Academic Press, p 67. Bierman, EL, and D Porte, Jr 1968. Carbohydrate intolerance and lipemia. Ann Int Med 68:926. Bierman, EL, JD Brunzell, RL Lerner, WR Hazzard, and D Porte, Jr 1970a. On the mechanisms of action of Atromid-S on triglyceride transport in man. Trans Assn Amer Phys 83:211. Bierman, EL, D Porte Jr, and JD Bagdade 1970b. Adipose tissue regulation and metabolic functions, edited by B Jeanrennaud and D Hepp. New York, Academic Press, p 209. Brunzell, JD, WR Hazzard, D Porte, Jr, and EL Bierman 1973. Evidence for a common saturable triglyceride removal mechanism for chylomicrons and very low density lipoproteins in man. J Clin Invest 52:1578. Brunzell, JD, D Porte, Jr, and EL Bierman 1971. Evidence for a common saturable removal system for dietary and endogenous triglyceride in man. J Clin Invest 50:15a. Caird, FI, A Pirie, and TG Ramsell 1969. Diabetes and the eye. Oxford and Edinburgh, Blackwell Scientific Publications, p 127. Catt, KJ 1970. Growth hormone. The Lancet i:933. Chochinov, RH, G Leslie, E Ullyot, and JA Moorhouse 1972. Sensory perception thresholds in patients with juvenile diabetes and their close relatives. New Eng J Med 286:1233. Chylack, LT, Jr , and JH Kinoshita 1969. A biochemical evaluation of a cataract induced in a high glucose medium. Invest Opthal 8:401. Clements, RS, Jr , and AI Winegrad 1969. Modulation of mammalian polyol:NADP oxidoreductase activity by ADP and ATP. Biochem Biophys Res Commun 36:1006. Clements, RS, Jr , and AI Winegrad 1972. Purification of alditol: NADP oxidoreductase from human placenta. Biochem Biophys Res Commun 47:1473. Cleve, H, K Alexander, HJ Mitzkat, P Nissen, and I Salzmann 1968. Serum glykoproteine beim diabetes mellitus; quantatitive immunologische Bestimmung von zaurem ¢;-Glycoprotein, G , apy-makroglobulin und Hamopexin bei diabetikern mit und Angiopathien. Diabetologia 4548. Daughaday, WH 1968. Textbook of endocrinology, edited by RH Williams. Philadelphia, Saunders, p 27. E Dixon, HBF 1972. A reaction of glucose with peptides. Biochem J 129:203. Eder, HA, NL Lasser, and HR Scholnick 1970. Lipid metabolism in diabetes mellitus, edited by M Ellenberg and H Rifkin. New York, McGraw-Hill, p 78. Eliasson, SG 1964. Nerve conduction changes in experimental diabetes. J Clin Invest 43:2353. Eliasson, SG 1969. Properties of isolated nerve fibres from alloxanized rats. J Neurol Neurosurg Psychiat 32:525. Engel, RME, DD Hoskins, and HG Williams-Ashman 1970. Enzymes of the nonphosphorylative (sorbitol) pathway for fructose biosynthesis of primate seminal vesicles. Invest Urol 7:333. Evans, JG, and LD Ostrander 1967. Fasting serum-triglyceride concentrations and distribution of subcutaneous fat. The Lancet 1:761. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. Demonstrable Metabolic Abnormalities 189 Fajans, SS 1971. What is diabetes? Definition, diagnosis, and course. Med Clins of N Amer 55:793. Feldman, EB, P Benkel, and RV Nayak 1963. Physiologic factors influencing circulating triglyceride concentrations in women: age, weight gain, and ovarian function. J Lab Clin Med 62:437. Fredrickson, DS, AM Gotto, Jr , and RI Levy 1972. The metabolic basis of inherited disease, edited by JB Stanbury, JB Wyngaarden, and DS Fredrickson. New York, McGraw-Hill, p 493. Fredrickson, DS, and RI Levy 1972. The metabolic basis of inherited disease, edited by JB Stanbury, JB Wyngaarden, and DS Fredrickson. New York, McGraw-Hill, p 545. Fredrickson, DF, K Ono, and LL Davis 1963. Lipolytic activity of post-heparin plasma in hyperglyceridemia. J Lipid Res 4:24. Friedenwald, JS and D Rytel 1955. Contribution to the histopathology of cataract. Arch Opthalmol 53:825. Gabbay, KH 1971. Development of neuropathy in streptozotocin diabetic rats. Diabetes 20 (Suppl 1):331. Gabbay, KH 1973. Vascular and neurological changes in early diabetes, edited by RA Camerini-Davalos and HS Cole. New York, Academic Press, p 417. Gabbay, KH, LO Merola, and RA Field 1966. Sorbitol pathway: presence in nerve and cord with substrate accumulation in diabetes. Science 151:209. Gertler, MM, HE Leetma, E Saluste, JJ Welsh, HA Rusk, DA Covalt, and J Rosenberger 1970. Carbohydrate, insulin, and lipid interrelationship in ischemic vascular disease. Geriatrics 25:134. . Glick, SM, J Roth, RS Yallow, and SA Berson 1965. The regulation of growth hormone secretion. Recent Progr Hormone Res 21:241. Glueck, CJ, RI Levy, and DS Fredrickson 1969. Immunoreactive insulin, glucose tolerance, and carbohydrate inducibility in Types II, III, IV, and V hyperlipoproteinemia. Diabetes 18:739. Goldstein, JL, WR Hazzard, HG Schrott, EL Bierman, and AG Motulsky 1972. Genetics of hyper- lipidemia in coronary heart disease. Trans Assn Amer Phys 85:120. Goldstein, JL, WR Hazzard, GH Schrott, EL Bierman, and AG Motulsky 1973a. Hyperlipidemia in coronary heart diseases. I. Lipid levels in 500 survivors of myocardial infarction. J Clin Invest 52:1533, Goldstein, JL, HG Schrott, WR Hazzard, EL Bierman, and AG Motulsky 1973b. Hyperlipidemia in coronary heart disease. II. Genetic analysis of lipid levels in 176 families and delineation of a new inherited disorder, combined hyperlipidemia. J Clin Invest 52:1544. Grace, CS, and RB Goldrick 1968. Fibrinolysis and body build. J Athero Res 8:705. Haust, MD 1970. Atherosclerosis, edited by RJ Jones. New York, Springer-Verlag, p 12. Havel, RJ. 1967. Vascular complications of diabetes Mellitus, edited by SJ Kimura, and WM Caygill. St. Louis, C. V. Mosby Co., p 114. Hayman, S, and JH Kinoshita 1965. Isolation and properties of lens aldose reductase. J Biol Chem 240:877. Hazzard, WR, JL Goldstein, HG Schrott, AG Motulsky, and EL Bierman 1973. Hyperlipidemia in coronary heart disease. III. Evaluation of lipoprotein phenotypes of 156 genetically defined survivors of myocardial infarction. J Clin Invest 52:1569. Hers, HG 1957. Le Métabolisme du Fructose. Bruxelles: Editions Arscia. 190 Diabetes Mellitus 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. Hers, HG 1960. L'aldose-reductase. Biochim Biophys Acta 37:120. Hollenberg, CH 1965. Adipose tissue, handbook of physiology, edited by AE Renold and GF Cahill, Jr. Washington, American Physiol Society, p 301. Hollister, LE, JE Overall, and HL Snow 1967. Relationship of obesity to serum triglyceride, cholesterol, and uric acid, and to plasma-glucose levels. Am J Clin Nutr 20:777. Holmquist, WR, and WA Schroeder 1966. A new n-terminal blocking group involving a Schiff base and hemoglobin Aer Biochemistry 5:2489. Kinoshita, JH 1965. Cataracts in galactosemia. Invest Opthalmol 4:786. Kinoshita, JH, GW Barber, LO Merola, and B Tung 1969. Changes in the levels of free amino acids and myoinositol in the galactose-exposed lens. Invest Opthalmol 8:625. Kinoshita, JH, D Dvornik, M Drami, and KH Gabbay 1968. The effect of an aldose reductase inhibitor on the galactose-exposed rabbit lens. Biochim Biophys Acta 158:472. Kipnis, DM 1970. Insulin secretion in normal and diabetic individuals. Advances in Int Med 16:103. Knopf, RF, SS Fajans, S Pek, JC Floyd, VK Prchkov, and JW Conn 1973. Vascular and neurological changes in early diabetes, edited by RA Camerini-Davalos, and HS Cole. New York, Academic Press, p 215. Kuck, JFR, Jr 1970a. Response of the mouse lens to high concentrations of glucose or galactose. Ophthal Res 1:166. Kuck, JFR, Jr 1970b. Biochemistry of the eye, edited by CN Graymore. New York, Academic Press, p 319. Kuwabara, T, JH Kinoshita, and DG Cogan 1969. Electron microscopic study of galactose-induced cataract. Invest Opthalmol 8:133. LeFevre, PG and RI Davis 1951. Active transport into the human erythrocyte: evidence from comparative kinetic and competition among monosaccharides. J Gen Physiol 34:515. Lundbaek, K 1971. Blood vessel disease in diabetes mellitus, edited by K Lundbaek and H Keen. Acta Diabetologica Latina, Vol VIII, Suppl 1, p 344. Lundbaek, K 1973. Vascular and neurological changes in early diabetes, edited by RA Camerini-Davalos, and HS Cole. New York, Academic Press, p 191. Lundbaek, K, NJ Christensen, VA Jensen, K Johansen, TS Olsen, AP Hansen, H Orskov, and R Osterby 1970. The pathogenesis of diabetic angiopathy and growth hormone. The Lancet 2:131. Mahler, R 1971. The effect of diabetes and insulin on biochemical reactions of the arterial wall. Acta Diabetologica Latina, Vol VIII, Suppl 1, p 68. Mahler, R 1973. Vascular and neurological changes in early diabetes, edited by RA Camerini- Davalos, and HS Cole. New York, Academic Press, p 49. Mayes, PA 1970. Adipose tissue regulation and metabolic functions, edited by B. Jeanrennaud and D Hepp. New York, Academic Press, p 186. McMillan, DE 1970. Changes in serum proteins and protein-bound carbohydrates in diabetes mellitus. Diabetologia 6:597. McMillan, DE 1972. Elevation of glycoprotein fucose in diabetes mellitus. Diabetes 21:863. Merimee, TJ 1973. Vascular and neurological changes in early diabetes, edited by RA Camerini- Davalos and HS Cole. New York, Academic Press, p 207. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 921. Demonstrable Metabolic Abnormalities 191 Merimee, TJ, SE Fineberg, VA McKusick, and JG Hall 1970a. Diabetes mellitus and sexual ateliotic dwarfism: A comparative study. J Clin Invest 49:1096. Merimee, TJ, MD Siperstein, JD Hall, and SE Fineberg 1970b. Capillary basement membrane structure: A comparative study of diabetics and sexual ateliotic dwarfs. J Clin Invest 49:2161. Molnar, GD, WF Taylor, and MM Ho 1972. Day-to-day variation of continuously monitored glycaemia: A further measure of diabetic instability. Diabetologia 8:342. Morrison, AD, RS Clements, Jr , SB Travis, F Oski, and AI Winegrad 1970. Glucose utilization by the polyol pathway in human erythrocytes. Biochem Biophys Res Commun 40:199. Morrison, AD, RS Clements, Jr, and AI Winegrad 1972. Effects of elevated glucose concentra- tions on the metabolism of the aortic wall. J Clin Invest 51:3114 Morrison, AD, and AI Winegrad 1973. Inhibition of polyol pathway ameliorates effects of elevated glucose levels in aorta (abstract). Clin Res 21:632. New, MI, TN Roberts, EL Bierman, and GG Reader 1963. The significance of blood lipid altera- tions in diabetes mellitus. Diabetes 12:208. Nikkila, EA 1969. Control of plasma and liver triglyceride kinetics by carbohydrate metabo- lism and insulin. Advances in Lipd Res 7:63. Nikkila, EA, TA Miettinen, MR Vesenne, and R Pelkonen 1965. Plasma-insulin in coronary heart disease. Response to oral and interavenous glucose and to tolbutamide. The Lancet 2:508. Nikkila, EA, K Pyorala, and M-R'Taskinen 1971. Role of insulinemia in arterial disease. Acta Diabetologica Latina, Vol VIII, Suppl 1, p 56. Patterson, JW 1953. Effect of lowered blood sugar on development of diabetic cataracts. Am J Physiol 172:77. Peters, N, and CN Hales 1965. Plasma insulin concentrations after myocardial infarction. The Lancet 1:1144. Pirie, A, and R van Heyningen 1964. The effect of diabetes on the content of sorbitol, glucose, fructose, and inositol in the human lens. Exptl Eye Res 3:124. Porte, D, Jr , and EL Bierman 1969. The effect of heparin infusion on plasma triglyceride in vivo and in vitro with a method for calculating triglyceride turnover. J Lab Clin Med 73:631. Reaven, GM, DB Hill, RC Gross, and JW Farquhar 1965. Kinetics of triglyceride turnover of very low density lipoproteins of human plasma. J Clin Invest 44:1826. Rifkind, BM, M Gale, and D Lawson 1968. Serum cholesterol and triglyceride levels and adiposity. Cardiovasc Res 2:143. Robertson, AL, Jr. 1968. Oxygen requirements of the human arterial intima in atherogenesis. Prog in Biochem Pharmacol 4:305. Robinson, DS 1965. Adipose tissue, handbook of physiology, edited by AE Renold, and GF Cahill, Jr. Washington, American Physiological Society, p 296. Rodman, HM, V Gupta, and BR Landau 1972. The polyol pathway in human placenta. Diabetes 21 (Suppl 1):329. Ross, R, and JA Glomset 1973. Atherosclerosis and the arterial smooth muscle cell. Science 180:1332, 192 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. Diabetes Mellitus Sailer, S, K Bolzano, F Sandhofer, P Spath, and H Braunsteiner 1968. Triglyceridspegel und Insulinkonzentration im Plasma nach oraler Glukosegabe bei Patienten mit primarer kohlenhydra- tinduzierter Hyperglyceridamie. Schweiz med Wschr 98c:1512. Segal, S 1972. The metabolic basis of inherited disease, edited by JB Stanbury, JB Wyngaarden, and DS Fredrickson. New York, McGraw-Hill, p 174. Spiro, RG 1963. Glycoproteins and diabetes. Diabetes 12:223. Spiro, RG 1969. Glycoproteins: Their biochemistry, biology, and role in human disease. New Eng J Med 281:991 (Part I), 1043 (Part II). Spiro, RG, and MJ Spiro 1971a. Studies on the biosynthesis of the hydroxylysine-linked disaccharide unit of basement membranes and collagens. I. Kidney glycosyltransferase. J Biol Chem 246:4899. Spiro, RG, and MJ Spiro 1971b. Studies on the biosynthesis of the hydroxylysine-linked disaccharide unit of basement membranes and collagens. II. Kidney galactosyltransferase. J Biol Chem 246:4910. Spiro, RG, and MJ Spiro 1971c. Studies on the biosynthesis of the hydroxylysine-linked disaccharide unit of basement membranes and collagens. III. Tissue and subcellular distribution of glucosyltransferases and the effects of various conditions on enzyme . levels. J Biol Chem 246:4919. Sterky, G, Y Larsson, and B Persson 1963. Blood lipids in diabetic and nondiabetic school children. Acta Paediat 52:11. Stewart, MA, MM Kurian, WR Sherman, and EV Cotlier 1968. Inositol changes in nerve and lens of galactose fed rats. J Neurochem 15:941. Stewart, MA, and JV Passoneau 1964. Identification of fructose in mammalian nerve. Biochem Biophys Res Commun 17:536. Stewart, MA, JV Passoneau, and OH Lowry 1965. Substrate changes in peripheral nerve during ischaemia and Wallerian degeneration. J Neurochem 12:719. Stewart, MA, WR Sherman, and S Anthony 1966. Free sugars in alloxan diabetic rat nerve. Biochem Biophys Res Commun 22:488. Stewart, MA, WR Sherman, MM Kurien, GI Moonsammy, and M Wisgerhof 1967. Polyol accumulations in nerve tissue of rats with experimental diabetes and galactosemia. J Neurochem 14:1057. Stout, RW 1968. Insulin stimulated lipogenesis in arterial tissue in relation to diabetes and atheroma. The Lancet 2:702. Stout, RW 1973. Vascular and neurological changes in early diabetes, edited by RA Camerini- Davalos, and HS Cole. New York, Acacemic Press, p 41. Tobian, LR, R Olson, and G Chesley 1969. Water content of arteriolar wall in renovascular hypertension. Am J Physiol 216:22. Travis SF, AD Morrison, RS Clements, Jr , AI Winegrad, and F Oski 1971. Metabolic alterations in the human erythrocyte produced by increases in glucose concentration: The role of the polyol pathway. J Clin Invest 50:2104. Trivelli LA, HM Ranney, and HT Lai 1971. Hemoglobin components in patients with diabetes mellitus. New Eng J Med 284:353. Tzagournis, M, R Chiles, JM Ryan, and TG Skillman 1968. Interrelationships of hyperinsulinism and hypertriglyceridemia in young patients with coronary heart disease. Circulation 38:1156. Tzagournis, M, JF Seidensticker, and GJ Hamwi 1967. Serum insulin, carbohydrate, and lipid abnormalities in patients with premature coronary heart disease. Ann Intern Med 67:42. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 123. Demonstrable Metabolic Abnormalities 193 Van Heyningen, R 1959. Formation of polyols by the lens of the rat with "sugar" cataract. Nature 184:194. Van Heyningen, R 1962. The sorbitol pathway in the lens. Exptl Eye Res 1:396. Van Heyningen, R 1968. The eye, 2nd ed., edited by H Davson. London, Academic Press, vol. I, chapter 6. Whereat, AF 1967. Atherosclerosis and metabolic disorder in the arterial wall. Exper Mol Pathol 7:233. Wick, AN, and DR Drury 1951. Action of insulin on the permeability of cells to sorbitol. Amer J Physiol 166:421. Williamson, JR 1973. Vascular and neurological changes in early diabetes, edited by RA Camerini-Davalos, and HS Cole. New York, Academic Press, p 225. Winegrad, AI, S Yalcin, and PD Mulcahy 1965. On the nature and treatment of diabetes, edited by BS Leibel and GA Wrenshall. Amsterdam, Excerpta Medica Foundation, p 452. Winzler, RJ 1964. Small blood vessel involvement in diabetes mellitus, edited by MD Siperstein, AR Colwell, Sr , and K Meyer. Washington, D.C., Am Inst of Biol Sci, p 243. Wolff, OH, and HB Salt 1958. Serum-lipids and blood-sugar levels in childhood diabetes. The Lancet 274:707. Wray, H, and AI Winegrad 1966. Free fructose in human cerebrospinal fluid. Diabetologia 2:82. Yalow, RS, SM Glick, J Roth, and SA Berson 1965. Plasma insulin and growth hormone levels in obesity and diabetes. Ann NY Acad Sci 131:357. Yde, H 1969. Abnormal growth hormone response to the ingestion of glucose in juvenile diabetics. Acta Med Scand 186:499. » 15 DIABETES, HYPERGLYCEMIA AND ATHEROSCLEROSIS: NEW RESEARCH DIRECTIONS Leon D. Ostrander, Jr. and Frederick H. Epstein A. THE ASSOCIATION BETWEEN DIABETES, GLUCOSE INTOLERANCE AND CARDIOVASCULAR DISEASE Atherosclerosis has been recognized as a complication of diabetes mellitus for more than a century (152), but the acute complications of ketoacidosis and infection commanded most attention until well into the insulin era. After the introduction of insulin, advances were rapid in both treatment and diagnosis. Tests of glucose tolerance (35,81,110) revealed that diabetes is a common chronic disease which often remains undiagnosed for may years (54). With expansion of the recognized diabetic population because of more sensitive diagnostic tests and virtual elimination of death from ketoacidosis or infection, chronic vascular disease has become the principal cause of death and disability among diabetics, with atherosclerotic cardiovascular disease accounting for most of the excess morbidity and mortality (45,90,103). The diabetic's tendency to develop gangrene of the feet has long been attributed to extensive atherosclerosis of the peripheral arteries, but neuropathy, microangiopathy, and susceptibility to infection may be as important (15). In other respects peripheral arterial disease does not differ appreciably from that of nondiabetics except for earlier onset, more extensive involvement of the vessels below the knees and a greater tendency for medial calcification (18,59,132,163). While peripheral arterial disease accounts for much morbidity among diabetics, mortality is due largely to atherosclerotic heart disease. Clinical studies (14,20,29,100,124), mortality statistics (45,90,103), and necropsy findings (18,32,69,144,145) attest to the diabetic's greater frequency and higher fatality rate from severe, premature coronary atherosclerosis. Women with diabetes appear to have nearly as high a prevalence of ischemic heart disease as diabetic men in necropsy and clinical studies. Nondiabetic women have a substantially lower in- cidence and prevalence of myocardial infarction than nondiabetic men. In a large autopsy series, Clawson and Bell (32) found that 19.5 percent of diabetic men and 17.4 percent of diabetic women past 40 years of age died of atherosclerotic heart disease. The proportions for nondiabetic men and women were 10.0 percent and 5.8 percent, respectively. Goldenberg and associates (69) found myocardial infarction at autopsy in 56 percent of men and 43.9 percent of women with diabetes, in contrast to 23.3 percent of men and 13.8 percent of women without known diabetes. Partamian and Bradley (124) reported 258 serial myocardial infarctions among patients from a diabetic clinic. Fifty-six percent occurred among women and 44 percent among men. The percentages correspond closely to the proportions of men and women in the base population. Evidence for a greater prevalence of cerebral vascular disease among diabetics is less con- vincing although suggestive of a relationship (18,34,60,101,114). Diabetes appears to be related to cerebral infarction but not to hemorrhage (6). Pathologic studies demonstrated earlier and more extensive cerebral artery atherosclerosis in diabetics than in nondiabetics (6,60,73). While cerebral infarction tends to occur at an older age than peripheral or coronary arterial manifesta- tions, it is an important cause of prolonged disability and ultimately death. 194 New Research Directions 195 An extensive literature indicates, therefore, that atherosclerosis is an important complica- tion of overt diabetes. In the past 25 years interest has gradually increased in the converse relationship, the frequency of abnormal carbohydrate tolerance among the very large number of persons with manifestations of atherosclerosis. Goldberger and associates (68) reported abnormal glucose tolerance in 10 to 14 patients with recent myocardial infarction. Waddell and Field (167) directed further attention to this area of research when they reported that 78 percent of 47 per- sons with severe manifestations of atherosclerosis were diabetic according to the cirteria of Fajans and Conn (54), while another 7 percent were probable diabetics. Other reports indicate that diminished glucose tolerance occurs in 33 percent to 46 percent of patients with healed myocardial infarction (5,156,168) and from 32 percent to 75 percent of persons with occlusive arterial disease in the legs (17,79,171). Although these observations suggest an excessive prev- alence of reduced glucose tolerance and, by inference, mild diabetes among persons with athero- sclerotic cardiovascular disease, appropriate control groups were not studied, the effect of age on glucose tolerance was not appreciated (8) and little consideration was given to the possibil- ity that the vascular event itself, particularly myocardial infarction, could influence glucose tolerance (41,98). Reaven and associates (137) reported a carefully controlled study of patients with healed myocardial infarction in which 41 percent had ‘abnormal glucose tolerance tests accord- ing to the criteria of Fajans and Conn (54), and an additional 34 percent had abnormal glucose tolerance tests after cortisone administration (53). Ten percent of controls had diabetes accord- ing to standard glucose tolerance tests, and another 23 percent had abnormal tests after cortisone In partial answer to criticisms that glucose intolerance may develop only after myocardial infarction, a few studies of other manifestations of ischemic heart disease revealed similar, significant associations with diabetes or specified levels of hyperglycemia (20,51,169). These clinical observations are supplemented and reinforced by epidemiological investigations based on defined populations. The Tecumseh epidemiological study was instituted to investigate the prevalence and incidence of chronic diseases and their precursors in a total, natural commun- ity. Participation of the residents approached 90 percent for each series of examinations. Cardiovascular disease and diabetes were studied with particular emphasis. Men and women with coronary heart disease, peripheral or cerebral vascular disease, or asymptomatic T wave inversion in the electrocardiogram had a significantly higher prevalence of hyperglycemia than the population as a whole (120). Hyperglycemia was defined as the upper quintile of age- and sex-specific blood glucose values. The relationship of hyperglycemia to cardiovascular disease was largely indepen- dent of blood pressure level or cholesterol concentration (49). Keen and associates (89) then reported similar observations from the epidemiological study in Bedford, England. There is substantial evidence that heart disease of any etiology, particularly when associated with congestive heart failure, may suppress insulin secretion and cause hyperglycemia (50,75,153). Therefore, reports of a higher than expected incidence of ischemic heart disease among hyper- glycemic but otherwise healthy participants in the Tecumseh and Bedford studies have strengthened the hypothesis that hyperglycemia is a true precursor of atherosclerotic heart disease (87,118). Stamler and associates (158) reported similar findings from their large epidemiological study of employed men in Chicago. A recent analysis of data from several epidemiological studies of cartiovascular disease, including that in Framingham, strongly suggests that hyperglycemia is an important precursor of atherosclerotic coronary artery disease (48). 196 Diabetes Mellitus The prevalence of hyperglycemia and its association with atherosclerosis differ greatly among certain ethnic groups. Diabetes and hyperglycemia are almost unknown among Eskimos (111), but diabetes according to clinical criteria is found in 50 percent of adult Pima Indians (22). Other ethnic groups have prevalence and incidence rates between these two extremes (62,142,143, 147,159). However, the frequency of atherosclerotic complications does not parallel the preva- lence of diabetes (21,61,14,147,172). The relative immunity of certain ethnic groups with a high prevalence of hyperglycemia and diabetes may be related to factors discussed in the next section. In summarizing this background information, the following conclusions appear justified: (1) Diabetics develop earlier and more extensive atherosclerosis of the coronary, peripheral, and cerebral arteries than nondiabetics. (2) Persons with atherosclerotic cardiovascular disease have a significantly higher prevalence of hyperglycemia than the general population. (3) Appar- ently healthy persons who develop ischemic heart disease have a significantly higher frequency of prior hyperglycemia than individuals who remain free of coronary events. B. FEATURES THAT PREDISPOSE HYPERGLYCEMIC AND DIABETIC PERSONS TO ATHEROSCLEROSIS Since overt diabetes and hyperglycemia predispose the individual to premature atherosclerosis, investigators have tried to identify characteristics of hyperglycemic persons which account for accelerated atherogenesis. Arterial hypertension, high serum lipid concentrations, hyperglycemia, increased coagulability of blood, and aberrations of insulin secretion are the major suspect con- ditions. I. Arterial Hypertension Much evidence suggests that diabetics have a higher prevalence of hypertension than nondiabet- ics (19,102,112,120,146). Pell and D'Alonzo (125) initially attributed most of the excess preva- lence of atherosclerotic heart disease observed in the diabetic employees of a large industrial concern to their higher frequency of hypertension. In a subsequent prospective mortality study, hypertensive diabetics were particularly prone to death but among normotensive employees, diabet- ics also had a substantially higher incidence of cardiovascular death than nondiabetics (126). Most investigators agree that hypertension increases the risk of atherosclerosis among diabetics, but some assign it a major role (69,97,166), while others find the relationship less convincing (86,101,132,144,170). There are many hypotheses but no definitive studies to account for the high prevalence of hypertension among persons with diabetes or hyperglycemia. II. Hyperlipidemia Hyperlipidemia in the form of gross hyperchylomicronemia is frequently observed in markedly insulin-deficient diabetics. It is due to inadequate postheparin lipolytic activity (PHLA), which is partially insulin dependent (12). Diabetics in ketoacidosis may have profound elevations of all lipid fractions which clear rapidly with insulin therapy (67). The defect is primarily defi- cient clearance of triglycerides because of inadequate insulin (13). Even among milder diabetics, delayed removal of triglycerides may be due to subtle deficiencies in PHLA (9,27,130). Among a much larger proportion of diabetics and hyperglycemic persons, high concentrations of triglyceride-rich, very low density lipoproteins (VLDL) are found without evidence of marked in- sulin insufficiency or any tendency to ketoacidosis (2,99,115,149,174). Albrink and associates (2) reviewed clinical records of all diabetics in a single medical center who had serum lipid determinations between 1931 and 1961. They observed that serum triglyceride concentrations in- New Research Directions 197 creased among diabetics during that period and were associated with a greater prevalence of ischemic heart disease in the decade 1951-1961. Higher triglyceride levels were attributed to liberalization of dietary carbohydrate allowances, but the authors conceded that more frequent obesity of patients in later decades may have influenced the results. Unfortunately, data on adiposity were incomplete. Regardless of etiology, numerous reports implicate hyperlipidemia, principally hypertriglyceridemia, as an important factor in the development of atherosclerosis among diabetics (33,55,76,78,99,122,149,164). However, Carlson and Wahlberg (30), using an in- travenous glucose tolerance test, could not demonstrate a significant relationship between glu- cose and lipid levels among patients with ischemic heart disease. Carbohydate induction of hypertriglyceridemia is a distinct entity (1,94), but Bierman and Porte (24) and Ford and associates (64) presented evidence that obesity is commonly related to high triglyceride and glucose concentrations. Bierman and Porte suggested that hyperglycemia and hypertriglyceridemia usually coexist because of a common etiologic factor, adiposity, without an independent metabolic relationship. Studies among apparently healthy men in a random sample of the Tecumseh population lend some support to this hypothesis (121), although the relationship may be different in the diabetic or hyperglycemic person (119). Reaven and associates (138) and Kuo and Feng (95) presented evidence that high serum insulin concentrations are common among maturity-onset diabetics and persons with hypertriglyceridemia. They postulated that insulin probably stimulates production of triglyceride-rich VLDL. Nikkila and Taskinen (116) pointed out the extreme complexity of interrelationships between serum insulin concentration, glucose toler- ance, adiposity and carbohydrate consumption and their association with hypertriglyceridemia. Most evidence relates adiposity to serum insulin concentration (11,85,127) so that identification of obesity as the most frequent cause of both hypertriglyceridemia and abnormal glucose tolerance is a useful generalization in spite of many exceptions (23). The relationship of hyperlipidemia to diabetes was further complicated by the introduction of the typing of plasma lipoprotein patterns to characterize lipid disorders. Lipoprotein types III, IV, and V of Fredrickson and associates (65) include high concentrations of VLDL and are usually associated with hyperglycemia. Differentiation between primary hyperlipoproteinemia with associated carbohydrate intolerance and diabetes with secondary hyperlipidemia is largely arbi-. trary when this system of classification is employed and the differentiation is usually based on the relative severity of the lipid and carbohydrate manifestations. Adiposity, delayed insulin responses after carbohydrate challenge and later hyperinsulinemia are features common to types III, IV, and V hyperlipoproteinemia and to maturity-onset diabetes (93,165), The approach to the study of inherited abnormalities of lipid metabolism associated with coronary heart disease was fundamentally altered by the recent studies by Goldstein et al. (70, 71) and the related publication by Hazzard et al. (77). These workers studied a large number of survivors of myocardial infarction and found by comparison with a large series of controls that 31 percent had hyperlipidemia; they then carried out studies of plasma triglyceride and choles- terol concentrations in the relatives and spouses of hyperlipidemic and normolipidemic survivors. The distribution of the fasting cholesterol and triglyceride values in the relative, together with segregation analysis, suggested the presence of five distinct lipid disorders. Three of these-- familial hypercholesterolemia, familial hypertriglyceridemia, and a newly defined disorder, familial combined hyperlipidemia--appeared to represent dominant expression of three different autosomal genes, occurring in about 20 percent of the survivors below 60 years of age and 7 per- cent of older survivors. Their data also suggested that heterozygosity for one of these three 198 Diabetes Mellitus disorders may have a frequency in the general population of about 1 percent, thus being among the most common inherited metabolic abnormalities. In a related study Hazzard et al. (77) demon- strated that on an individual basis no lipoprotein pattern proved to be specific for any partic- ular genetic lipid disorder, and that conversely in the population studied no genetic disorder was specific by a single lipoprotein pattern. Their studies indicated that lipoprotein pheno- types are not qualitative markers in the genetic sense, but are quantitative parameters which may vary among individuals with the same genetic disorder. The implication is that, for the moment, the genetic classification of the individual hyperlipidemic patient may be more meaningfully ap- proached from a quantitative analysis of lipid levels in his relatives. In a follow-up, Brunzell et al. (28) examined the prevalence of diabetes mellitus in 397 adult first degree relatives of 91 index subjects with an autosomal dominant form of hypertriglyceridemia (familial hyper- triglyceridemia or familial combined hyperlipidemia) which had been established by family studies. This study demonstrated that diabetes mellitus and these common familial forms of hypertriglyc- eridemia segregate independently, for they could find no evidence that the prevalence of diabetes mellitus was related to the presence of a specific inherited form of hypertriglyceridemia. They suggested that the frequently reported association between diabetes mellitus and hypertriglyc- eridemia may reflect preferential selection of patients with both diabetes mellitus and a familial form of hypertriglyceridemia, since the combination is more likely to be symptomatic than hypertriglyceridemia alone. III. Abnormalities Related to Coagulation and Thrombosis There are no convincing reports of an excessive incidence of thrombotic or thromboembolic events related to uncomplicated diabetes (25,66,105), although several aspects of coagulation are frequently abnormal in diabetics. Increased platelet adhesiveness is the most commonly reported abnormality (10,26,96,106,136,154). High concentrations of free fatty acids, a common finding in diabetes (135), causes platelet aggregation (36,37,80), but several studies suggest that in- creased platelet adhesiveness among diabetics is independent of fatty acid level (96,117). One report attributes platelet '"stickiness' directly to the blood glucose concentration (26), but this observation has not been confirmed. Diabetics have higher concentrations of factor VIII, factor V, and fibrinogen than nondiabetics (43,106,136). Diabetics also appear to have lower than normal fibrinolytic activity (10,56). Hypercoagulability, particularly increased platelet adhesiveness, has long been implicated as a factor in atherosclerosis among both diabetics and nondiabetics (113,128). Connor and as- sociates (37) demonstrated that high concentrations of long chain saturated fatty acids cause marked platelet aggregation and massive thrombosis in experimental animals; they postulated that lower concentrations of the same fatty acids probably initiated enough platelet aggregates to form small mural thrombi, which in turn became sites of atheroma formation. Although still some- what conjectural, platelet thrombi are probably a factor in atherogenesis. Diabetics are then at higher risk of atherosclerosis than nondiabetics in part because of abnormalities related to coagulation and thrombosis. IV. Insulin and Blood Glucose Concentrations Inadequate or excessive insulin release, inappropriate response to insulin, and abnormal in- sulins are suspected factors in the initiation of premature atherosclerosis in diabetes. Winegrad and associates, whose studies are described elsewhere in this monograph, relate the New Research Directions 199 formation of polyols in the inner arterial wall directly to hyperglycemia and hypoinsulinemia (175). Sorbitol, the principal polyol formed in aorta, alters metabolism of the inner arterial wall and increases its susceptibility to atheroma formation. Martin and Stocks (104) reported a significant relationship between atherosclerosis and a reduced hypoglycemic effect from intravenous insulin injections among first insulin dependent and later less severe diabetics (161). Sloan and associates (155) reported higher serum insulin re- sponses and blood glucose levels after glucose challenge among apparently nondiabetic men with atherosclerotic peripheral vascular disease than among age-matched overtly healthy control sub- jects. On the other hand, Elkeles and associates (44) observed more frequent vascular complica- tions among noninsulin dependent diabetics with poor endogenous insulin responses than in sub- jects with more appropriate insulin levels after oral glucose stimulation. Diabetes was long attributed to hypoinsulinemia due to inadequate islet cell mass or func- tion. This concept logically explains the pathophysiology of most insulin-dependent and many less severe diabetics (52,107,176). The maximum serum insulin response after stimulation occurs ap- preciably later in most diabetics than in nondiabetics, although the peak insulin concentration may be quite high (63,91,107,151). Some persons who are diabetic according to glucose tolerance criteria react promptly to stimulation with either normal or supernormal insulin responses (40, 83,140,139). Hyperglycemia in the presence of apparently adequate serum insulin concentra- tions suggests either a high proportion of biologically inactive insulin or resistance to the ef- fect of insulin. Obesity is the most frequent cause of hyperinsulinemia because of the high con- centrations required to overcome resistance to glucose transport into large fat cells (148). Glucocorticoids, glucagon, and growth hormone are insulin antagonists, but high concentrations of these hormones are probably not common factors in the etiology of diabetes or atherosclerosis. Catecholamines cause gluconeogenesis, a factor in hyperglycemia, and have been implicated in the development of heart disease (74,133,134). Epinephrine and norepinephrine also induce hyper- glycemia by inhibiting insulin secretion (31,129), so that adrenergic excess results in hyper- glycemia but, at the same time, hypoinsulinemia. Therefore, while hyperinsulinemia in nonobese diabetics may be due to insulin resistance, the mechanism is unknown. The variety of glucose-insulin relationships observed among diabetics is inconsistent with a single pathophysiologic explanation. Furthermore, individuals may exhibit markedly different glucose or insulin responses when studied under similar conditions over periods of days, months, or years (52). Although a single hypothesis cannot reconcile all experimental observations, one can generalize that all diabetics have inappropriate insulin-glucose relationships, which are frequently unstable over time. Not only hyperglycemia but hyperinsulinemia may enhance athero- sclerosis. Stout (162) reported that high concentrations of serum insulin increase lipid syn- thesis in the arterial intima and inner media. Epstein (47) earlier hypothesized that either insulin deficiency or excess could enhance atherosclerosis. C. TREATMENT 1. Medical Treatment of Hyperglycemia Blood glucose levels as close to normal as possible have been the logical but usually un- attainable goal of diabetic management since before the insulin era. The frequent occurrence of both microangiopathy and premature atherosclerosis among persons with only mild abnormalities of glucose tolerance or well controlled overt diabetes suggests that factors other than meticulous 200 Diabetes Mellitus control of the glucose concentration are important (51,87,169). However, because of the chron- icity of diabetes and the variability of glucose levels over long periods, a major effect of hyperglycemia cannot be discounted in spite of apparent stability of glucose concentrations during a particular interval of time. The University Group Diabetes Program (92), a long-term, randomized, prospective inter- vention study, treated noninsulin dependent diabetics according to five regimens: fixed dose insulin, variable dose insulin, tolbutamide, phenformin, and placebo. The results continue to evoke controversy (39,58,108,131,150), but much of it concerns the significantly higher cardio- vascular mortality in the tolbutamide treated group. The most important observation was the failure of any treatment regimen to reduce mortality from atherosclerotic events below the rate of the’ placebo group. It is not surprising that no regimen of treatment reduced the incidence of atherosclerotic cardiovascular events. Control, as determined by periodic serum glucose concentrations, was appreciably better only in the variable dose insulin group (108). Adherence to the program was lowest, only 45.3 percent, in this group. Perhaps members of the variable dose insulin segment achieved their mean 13 percent reduction in fasting glucose concentration by more regular insulin administration prior to clinic visits. Mean serum glucose levels prob- ably varied little among any of the study groups. The study design did not include standardized treatment of other precursors of athero- sclerotic events, but evidence of vascular disease, renal insufficiency and rather marked hyper- cholesterolemia was taken into account in the characterization of the groups (92). The University Group Diabetes Program participants were a heterogeneous collection of dia- betics whose antecedents of macroangiopathy were incompletely defined and essentially untreated even when recognized. Consequently, the study revealed neither the relative importance of separate precursors of atherosclerosis nor the benefits, if any, of intervention. Keen and associates (88) found no evidence of a harmful effect from tolbutamide in a pri- mary intervention study among ''borderline' diabetics. Lower morbidity from ischemic heart dis- ease was observed among the treated participants, but the difference from the control group was not statistically significant. Paasikivi (123) reported that long-term tolbutamide treatment after first myocardial infarction postponed fatal and nonfatal recurrences among persons with abnormal intravenous glucose tolerance tests but without overt diabetes. The subjects of the three intervention studies were not comparable. The inconsistent results of tolbutamide treat- ment on atherosclerotic complications were probably due to subject selection. The effect of tolbutamide, whether favorable or deleterious, appears minor and inconclusive. Available evi- dence suggests that no current regimen to control blood glucose concentration has an appre-= ciable effect on the atherosclerotic complications of diabetes. II. Surgical Treatment of Atherosclerotic Complications Surgeons vary in their optimism regarding the results of operations to alleviate peripheral arterial insufficiency in diabetics. Barker (15) makes a clear distinction between occlusive disease in diabetics that is similar to that found in nondiabetics and the more complicated vascular lesions associated with rest pain, gangrene, diffuse small vessel disease and neurop- athy. He considers patients with the latter findings poor operative risks with minimal prospects for improvement with any form of treatment. Most surgeons agree that diabetics tend to have more extensive atherosclerosis of the tibial and peroneal arteries than nondiabetics and more frequent New Research Directions 201 medial calcification (15,16,38,173). Diabetics are operated upon more often for relief of rest pain or treatment of gangrene than nondiabetics. The preferred revascularization procedure is the saphenous vein bypass, which is said to be as effective in diabetics as in nondiabetics for relief of discrete occlusions (16,160,173). Even in patients with gangrene, revascularization often permits a more conservative amputation. In selected cases, sympathectomy and endarterectomy are still useful, but prosthetic tubes are no longer used extensively because of thrombosis, which is particularly common below the knee (15). At best, surgery is only palliative treatment for peripheral atherosclerosis in diabetes. Aortocoronary saphenous vein bypass for treatment of atherosclerotic occlusive disease of the coronary arteries has gained popularity in recent years. Only limited information is avail- able on the short-term results of operation and no assessment is available of large numbers of patients observed for 2 years of more. Immediate results from several centers are encouraging. Sixty to ninety percent of survivors obtain complete to moderate relief from intractable angina pectoris (4,42,109). The surgical mortality is about 5 percent and another 10 percent survive myocardial infarction as a complication of the operation. No distinction has been made between diabetics and nondiabetics in published reports. Poor results have generally been attributed to graft occlusion or inadequate myocardial function preoperatively (4,7,141). Lack of adequate follow-up and failure to perform controlled studies of the efficacy of the procedure make objec- tive evaluation impossible at this time (157). It is probably a worthwhile palliative procedure for some patients but is not a proven treatment for occlusive disease of the coronary arteries. D. RESEARCH NEEDS I. Biological Definition of Hyperglycemic States Since no biological marker is specific for diabetes, diagnosis depends on arbitrary criteria for classification of the glucose tolerance test, the insulin response to specific stimuli, and the presence of other clinical features that are characteristic of diabetes in varying degrees. Diabetes may represent the upper end of a continuous distribution of glucose tolerance and other variables; specific biological identification is then impossible. Further productive research on prevention or delay of atherosclerotic complications does not depend on a diagnostic marker of diabetes. Diabetes in the broadest sense should be classified in a stratified fashion according to well-defined criteria that include: (a) blood glucose concentrations during the glucose toler- ance test, (b) serum insulin concentrations at appropriate intervals after stimulation, and (c) levels of associated factors of suspected importance in the development of atherosclerotic complications (age, sex, adiposity, blood pressure, serum lipids, smoking, and coagulation factors). Classification would depend on a comprehensive examination and at least two and preferably three successive measurements of physiological variables and biochemical factors. It is inconceivable that maturity-onset diabetics, who differ substantially in the features enumer- ated, would exhibit similar biochemical interrelationships or experience the same risk of atherosclerotic complications. Therefore, multifactorial classification and data analysis are essential for further investigation of atherosclerosis in diabetes. II. Clinical and Laboratory Research Much additional information is needed to develop a comprehensive concept of the multiple 202 Diabetes Mellitus complex metabolic interrelationships between blood glucose, serum insulin, lipids, blood pressure, and coagulation factors in persons with different degrees of carbohydrate intolerance and various expressions of the other enumerated factors. Complex problems in intermediary metabolism would hopefully be clarified if not entirely settled by more precise classification of subjects. Human and animal studies should be designed to determine environmental and genetic factors that predispose subjects with various glucose-insulin relationships to hypertension, hyperlipid- emia, and coagulation abnormalities. Meticulous study of humoral and biochemical dynamics under strictly controlled experimental conditions should yield essential information about metabolic interrelationships and their correlation with atherosclerosis. Angioradiography will probably be an important adjunct to such research in humans, because it permits accurate assessment of sub- clinical arterial atherosclerosis. With appropriate precautions and skillful technique, risk to the subject is minimal, and arteriograms allow immediate correlation between metabolic features and degree of atherosclerosis in critical arteries. III. Epidemiological Studies Sufficiently large numbers of otherwise healthy hyperglycemic subjects with various combina- tions of glucose and insulin concentrations and other suspect characteristics and an appropriate control group must be followed for an adequate period of time to ensure that enough atheroscle- rotic events occur among the cohort for valid statistical determination of relative risk. Par- ticipants should range from 35 through 59 years of age, include approximately equal numbers of men and women, and all be free of evident cardiovascular disease. They should be geographically stable and willing to participate for 10 years. Approximately 500 myocardial infarctions or coronary deaths would be necessary to assess adequately the suspected precursors. It is esti- mated that population of 5,000 to 7,000 diabetics would be required in order to collect such a large number of definite atherosclerotic events within 10 years. The larger number is probably more realistic if one allows for attrition from causes other than coronary events. Such a study requires collaboration among investigators in many centers. Although the logistics and cost are formidable, such a study should yield information that more than justifies the expense. Particular features or combinations of traits probably account for a disproportionately large segment of atherosclerotic events among hyperglycemic persons. Only a well-designed epidemiological study can. test this hypothesis. IV. Intervention Studies The design and initiation of precise intervention studies require prior identification of specific precursors of atherosclerosis among different categories of hyperglycemics. Problems of design, population cooperation, and cost would be great, but the causative role of traits implicated as precursors of atherosclerosis in clinical, laboratory and epidemiological studies can only be established if successful intervention reduces the incidence of the disease. V. Early Identification of Potential Diabetics Individuals may have features of diabetes other than hyperglycemia for long periods before glucose intolerance is readily detectable (54). Sufficient numbers of such subclinical diabet- ics could be included in a prospective epidemiological study to ascertain their relative risk of atherosclerotic events. Further detailed study of the intermediary metabolism of such per- sons should elucidate more of the hormonal and biochemical interrelationships of early diabetes. New Research Directions 203 VI. Investigation of Ethnic Groups with a High Prevalence of Diabetes but Little Atherosclerosis Intensive laboratory and clinical investigation of the physiological features and inter- mediary metabolism of appropriate samples of such populations should afford important clues to the origins of atherosclerotic complications in the general diabetic population. VII. Summary Classification of hyperglycemic persons according to the presence and severity of major factors of suspected importance in the development of atherosclerosis is essential to the proposed investigations. This research is based on the hypothesis that atherosclerosis is a multifactorial disease, in which the etiologic importance of abnormal carbohydrate and insulin metabolism has not been fully appreciated. The suggested epidemiological studies require standardized procedures and research sophistication far beyond any previously attempted among large numbers of partici- pants. However, less ambitious studies are unlikely to yield enough new information to justify subsequent intervention trials or prevention programs. E. APPLICATION OF PROPOSED RESEARCH TO PREVENTIVE MEDICINE Complications of atherosclerosis are the leading cause of death and disability in the United States (57). Atherosclerotic events are not confined to the elderly but are the major cause of mortality and morbidity among middle-aged employed men (72). Current evidence suggests that atherosclerosis is caused by combinations of predisposing conditions (84). Epidemiological studies have revealed such a strong and consistent association between serum cholesterol level, arterial blood pressure, and cigarette smoking, singly and in combination, and the incidence of atherosclerotic heart disease, that they are generally accepted as .important precursors (46,82). While the evidence is less extensive, diabetes and hyperglycemia are also significantly related to the risk of ischemic heart disease (48,87,118,158). The interrelationships of these pre- cursors have been discussed at length in this article. A clear categorization of diabetes, hyperglycemia, and associated conditions, investigation of basic hormonal and biochemical relationships, and determination of the relative risk of spe- cific combinations of features in each category would provide the essential base for the rational application of prophylaxis against atherosclerosis. Much can be surmised from current knowledge, but the problem requires more precise definition before extensive preventive programs are feasi- ble. Even a modest reduction in the incidence of atherosclerotic events by the application of information derived from the research suggested would more than justify the considerable effort and expense. REFERENCES 1. Ahrens, EH, Jr, J Hirsch, K Oette, JW Farquhar, and Y Stein 1961. Carbohydrate-induced and fat-induced lipemia. Trans Assoc Amer Phys 74:134-146. 2. Albrink, MJ, PH Lavietes, and EB Man 1963. Vascular disease and serum lipids in diabetes mellitus. Observations over thirty years (1931-1961). Ann Int Med 58:305-323. 3. Albrink, MJ, and EB Man 1958. Serum triglycerides in health and diabetes. Diabetes 7:194-200. 4. Alderman, EL, HJ Matlof, L Wexler, NE Shumway, and DC Harrison 1973. Results of direct coronary artery surgery for the treatment of angina pectoris. New Eng J Med 288:535-539. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 204 Diabetes Mellitus Aleksandrow, D, M Ciswicko-Szanjdermon, H Tgnatowska, and B Wocial 1962. Studies on disturbances of carbohydrate metabolism in atherosclerosis. J Ather Res 2:171-180. Alex, M, EK Baron, S Goldenberg, and HT Blumenthal 1962. An autopsy study of cerebrovas- cular accident in diabetes mellitus. Cir 35:663-673. Anderson, RP 1972. Effects of coronary bypass graft occlusion on left ventricular perfor- mance. Circ 46:507-513. Andres, R 1971. Aging and diabetes. Med Clinics N Amer 55:835-846. Avogaro, P, C Capri, G Cazzolato, and M Pais 1972. Alterazioni lipidemiche nel diabete mellito. Acta Diabetica Latina 9:540-561. Badawi, H, M El1-Sawy, M Mikhail, AM Nomeir, and S Tewfik 1970. Platelets, coagulation and fibrinolysis in diabetic and nondiabetic patients with quiescent heart disease. Angiology 21:511-519. Bagdade, JD, EL Bierman, and D Porte, Jr 1967b. The significance of basal insulin levels in the evaluation of the insulin response to glucose in diabetic and nondiabetic subjects. J Clin Invest 46:1549-1557. Bagdade, JD, D Porte, Jr, and EL Bierman 1967a. Diabetic lipemia: A form of acquired fat- induced lipemia. New Eng J Med 276:427-433. Bagdade, JD, D Porte, Jr, and EL Bierman 1968. Acute insulin withdrawal and the regulation of plasma triglyceride removal in diabetic subjects. Diabetes 17:127-132. Bailey, RR, and DW Beaven 1968. Diabetes mellitus and myocardial infarction. Australasian Ann Med 17:312-314. Barker, WF 1971. Peripheral vascular disease in diabetes. Diagnosis and management. Med Clinics N Amer 55:1045-1055. Barner, HB, GC Kaiser, and VL Willman 1971. Blood flow in the diabetic leg. Circ 43: 391-394. Bartels, CC, and FR Rullo 1958. Unsuspected diabetes mellitus in peripheral vascular disease. New Eng J Med 259:633-635. Bell, ET 1952. A postmortem study of vascular disease in diabetics. Arch Path 53:444-455. Bell, ET, and BJ Clawson 1928. Primary (essential) hypertension. Arch Pat 5:939-1002. Bellet, S, and L Roman 1967. The exercise test in diabetic patients as studied by radio- electrocardiography. Cir 36:245-254. Bennett, PH 1973. Prevalence of ECG changes in Pima Indians and the population of Tecumseh, Michigan. In Vascular and neurological changes in early diabetes. Eds, RA Camerini-Davalos, and HS Cole. New York, Academic Press, pp 39-40. Bennett, PH, TA Burch, and M Miller 1971. Diabetes mellitus in American (Pima) Indians. Lancet 2:125-128. Bierman, EL 1973. Hypertriglyceridemia in early diabetes. In Vascular and neurological changes in early diabetes. Eds, RA Camerini-Davalos, and HS Cole. New York, Academic Press, pp. 67-72. Bierman, EL, and D Porte, Jr 1968. Carbohydrate intolerance and lipemia. Ann Int Med 68:926-933. Borhani, N 1972. Endocrine factors and venous thrombosis. Mil Mem Fund Quart 50: (part 2) 31-45. Bridges, JM, AM Dalby, JHD Millar, and JA Weaver 1965. An effect of d-glucose on platelet stickiness. Lancet 1:75-77. 27. 28. 29. 30. 31. 32. 33. 34. 3s. 36. 37. z8. 39. 40. 41. 42. 43. 44. 45. 46. 47. Brunzell, JD, D Porte, Jr, and EL Bierman 1971. system for dietary and endogenous triglyceride in man (abstract). Brunzell, JD, WR Hazzard, AG Motulsky, and EL Bierman 1974. hypertriglyceridemia (abstract). Bryfogle, JW, and RF Bradley 1957. betes 6:159-167. Carlson, LA, and F Wahlberg 1966. interrelation studied in ischemic cardiovascular disease. 307-315. Cerasi, E, R Luft, and S Efendie 1971. A dose-response study in man. insulin secretion. Clawson, BJ, and ET Bell 1948. diabetic persons. New Research Diabetes 23 (Supplement 1):351. Directions 205 Evidence for a common saturable removal J Clin Invest 50:15a. Prevalence of diabetes in The vascular complications of diabetes mellitus. Dia- Serum lipids, intravenous glucose tolerance and their Arch Path 48:105-106. Cohen, LS, WC Elliott, MD Klein, and R Gorlin 1966. Coronary heart disease. cinearteriographic and metabolic correlations. Amer J Card 17:153-168. Conant, RG, JA Perkins, and AB Ainley 1965. potential. J Chron Dis 18:397-403. Conn, JW 1940. Interpretation of glucose tolerance test. Connor, WE 1962. Connor, WE, JC Hoak, and ED Warner 1963. Acta Medica Scandinavica 180: Antagonism between glucose and epinephrine regarding Acta Med Scand 190:411-417. Incidence of fatal coronary disease in nondiabetic and in Clinical, Stroke morbidity, mortality and rehabilitative Amer J Med Sci 199:555-564. The acceleration of thrombus formation by certain fatty acids. J Clin Invest 41:1199-1205. J Clin Invest 42:860-866. Conrad, MC 1967. severe vascular disease. Cornfield, J 1971. of the mortality findings. Massive thrombosis produced by fatty acid infusion. Large and small artery occlusion in diabetics and nondiabetics with The University Group Diabetes Program. J Amer Med Assoc 217:1676-1687. Circ 36:83-91. A further statistical analysis Danowski, TS, YB Lombardo, LV Mendelsohn, DG Corredor, CR Morgan, and G Sabeh 1969. Insulin patterns prior to and after onset of diabetes. Metabolism 18:731-740. Datey, KK, and NC Med 276:263-265. Effler, DB, RG Favaloro, and LK Groves 1970. Nanda 1967. Hyperglycemia after acute myocardial infarction. New Eng J vein graft techniques: clinical experience with 224 operations. J Thoracic Surg 59:147-154. Egebert, 0 1963. Lab Invest 15:533- Elkeles, RS, C Lowy, ADH Wyllie, JL Young, and TR Fraser 1971. The blood coagulability in diabetic patients. 538. Coronary artery surgery utilizing saphenous and Cardiovas Scandinavian J Clin and Serum insulin, glucose and lipid levels among mild diabetics in relation to incidence of vascular complications. Lancet 1:880-883. Entmacher, PS, HF Root, and HH Marks 1964. Diabetes 13:373-377. Epstein, FH 1965a. 18:735-774. Epstein, FH 1967. 619. The epidemiology of coronary heart disease. A review. Hyperglycemia. A risk factor in coronary heart disease. Longevity of diabetic patients in recent years. J Chron Dis Circ 36:609- vr 206 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 39. 60. 61. 62. 63. 64. 65. 66. Diabetes Mellitus Epstein, FH (in press). Glucose intolerance and coronary heart disease incidence--recent observations. In International symposium on lipid metabolism, obesity, and diabetes mellitus: Impact on atherosclerosis in hormone and metabolic research. Ed, EF Pfeiffer. Thieme Verlag, Stuttgart. Epstein, FH, LD Ostrander, Jr, BC Johnson, MW Payne, NS Hayner, JB Keller, and T Francis, Jr 1965b. Epidemiological studies of cardiovascular disease in a total community--Tecumseh, Michigan. Ann Int Med 62:1170-1187. Ettinger, PO, HA Oldewurtel, AB Weisse, and TJ Regan 1968. Diminished glucose tolerance and immunoreactive insulin response in patients with nonischemic cardiac disease. Circ 38:559- 567. Fabrykant, M, and ML Gelfand 1964. Symptom-free diabetes in angina pectoris. Amer J Med Sci 247:665-668. Fajans, SS 1971. What is diabetes: Definition, diagnosis, and course. Medical Clinics of North America 55:793-805. Fajans, SS, and JW Conn 1954. The approach to the prediction of diabetes mellitus by modification of the glucose tolerance test with cortisone. Diabetes 3:296-304. Fajans, SS, and JW Conn 1959. The early recognition of diabetes mellitus. Ann N Y Acad Sci 82:208-218. Falsetti, HL, JD Schnatz, DG Greene, and IL Bunnell 1968. Lipid and carbohydrate studies in coronary artery disease. Circ 37:184-191. Fearnley, GR, R Chakrabarti, and PRD Avis 1963. Blood fibrinolytic activity in diabetes mellitus and its bearing on ischemic heart disease and obesity. Brit Med J 1:921-923. Feinleib, M, and MJ Davidson 1972. Coronary heart disease mortality. A community perspective. J Amer Med Assoc 222:1129-1134. Feinstein, AR 1971. An analytical appraisal of the University Group Diabetes Program (UGDP) Study. Clin Pharmac and Therap 12:167-191. Ferrier, TM 1963. Radiologically demonstrable arterial calcification in diabetes mellitus. Australasian Ann Med 13:222-228. Flora, GC, AB Baker, RB Loewenson, and AC Klassen 1968. A comparative study of cerebral atherosclerosis in males and females. Circ 38:859-869. Florey, CD, H McDonald, J McDonald, and WE Miall 1972. The prevalence of diabetes in a rural population of Jamaican adults. Internat'l J Epid 1:157-166. Florey, CD, RDG Milner, and WE Miall 1972. Insulin excess as the initial lesion in diabetes. Lancet 2:227. Floyd, JC, Jr, SS Fajans, JW Conn, S Pek, and RF Knopf 1970. Subnormal secretion of insulin in mild diabetes mellitus. In Early diabetes. Eds, RA Camerini-Davalos and HS Cole. New York, Academic Press, pp 113-118. Ford, S, Jr, RC Bozian, and HC Knowles 1968. Interactions of obesity and glucose and insulin levels in hypertriglyceridemia. Amer J Clin Nutr 21:904-910. Frederickson, DS, RI Levy, and RS Lees 1967. Fat transport in lipoproteins: An integrated approach to mechanisms and disorders. New Eng J Med 276:32-44, 94-103, 148-156, 215-226, 273-281. Freiman, DG 1969. Venous thromboembolic disease in medical and malignant states. In Thrombosis. Eds, S Sherry, KM Brinkhous, E Genton, and JM Strengle. Washington, D.C., National Academy of Sciences, pp 5-18. 67. 68. 69. 70. 71% 72. 73, 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. New Research Directions 207 Gofman, JW, O Delalla, F Glazier, NK Freeman, FT Lindgren, AV Nichols, B Strisower, and AR Tamplin 1954. The serum lipoprotein transport system in health, metabolic disorders, atherosclerosis and coronary artery disease. Plasma 2:413-484. Goldberger, E, J Alesio, and F Woll 1945. The significance of hyperglycemia in myocardial infarction. N Y J Med 45:391-393. Goldenberg, S, M Alex, and HT Blumenthal 1958. Sequellae of atherosclerosis of the aorta and coronary arteries. A statistical study in diabetes mellitus. Diabetes 7:98-108. Goldstein, JL, WR Hazzard, HG Schrott, EL Bierman, and AG Motulsky 1973a. Hyperlipidemia in coronary heart disease. I. Lipid levels in 500 survivors of myocardial infarction. J Clin Invest 52:1533-1543. Goldstein, JD, HG Schrott, WR Hazzard, EL Bierman, and AG Motulsky 1973b. Hyperlipidemia in coronary heart disease. II. Genetic analysis of lipid levels in 176 families and delineation of a new inherited disorder, combined hyperlipidemia. J Clin Invest 52:1544- 1568. Gordon, T and WB Kannel 1971. Premature mortality from coronary heart disease. The Framingham Study. J Amer Med Assoc 215:1617-1625. Grunnet, ML 1963. Cerebrovascular disease: diabetes and cerebral atherosclerosis. Neurology 13:486-491. Haft, JI, PD Krans, FJ Albert, K Fani 1972. Intravascular platelet aggregation in the heart induced by norepinephrine. Microsopic studies. Circulation 46:698-708. Hait, G, M Corpus, FR Lamarre, S Yuan, J Kypson, and G Cheng 1972. Alteration of glucose and insulin metabolism in congenital heart disease. Circ 46:333-346. Hatch, FT, PK Reissell, TMW Poon-King, GP Canellos, RS Lees, and LM Hagopian 1966. A study of coronary heart disease in young men. Characteristics and metabolic studies of the patients and comparison with age-matched healthy men. Circ 33:679-703. Hazzard, WR, JL Goldstein, HG Schrott, AG Motulsky, and EL Bierman 1973. Hyperlipidemia in coronary heart disease. III. Evaluation of lipoprotein phenotypes of 156 genetically defined survivors of myocardial infarction. J Clin Invest 52:1569-1577. Heinle, RA, RI Levy, DS Frederickson, and R Gorlin 1969. Lipid and carbohydrate abnormal- ities in patients with angiographically documented coronary artery disease. Amer J Card 24:178-186. Hines, EA, Jr, and NW Barker 1940. Arteriosclerosis obliterans: a clinical and pathological study. Amer J Med Sci 200:717-730. Hoak, JC, ED Warner, and WE Connor 1967. Platelets, fatty acids and thrombosis. Circ Res 20:11-17. Hopkins, AH 1915. Studies in the concentration of blood sugar in health and disease as determined by Bang's micromethod. Amer J Med Sci 149:254-267. Inter-Society Commission for Heart Disease Resources 1970. Primary Prevention of the Atherosclerotic Diseases. Circ 42:A55-A95. Jackson, WPU, W van Mieghem, and P Keller 1972. Insulin excess as the initial lesion in diabetes. Lancet 1:1040-1044. Kannel, WB, TR Dawber, A Kagan, N Revotskie, and J Stokes III 1961. Factors of risk in the development of coronary heart disease--six-year follow-up experience. The Framingham Study. Ann Int Med 55:33-50. Karam, JH, GM Grodsky, and PH Forsham 1963. Excessive insulin response to glucose in obese subjects as measured by immunochemical assay. Diabetes 12:197-204. 208 Diabetes Mellitus 86. 37. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. Keen, H 1972. Personal communication. Keen, H, and RJ Jarrett 1970. The effect of carbohydrate tolerance on plasma lipids and atherosclerosis in man. In Atherosclerosis: Proceedings of the second international symposium. Ed, RJ Jones. New York, Heidelberg, Berlin, Springer-Verlag, pp 435-444. Keen, H, RJ Jarrett, JD Ward, and JH Fuller 1973. Borderline diabetics and their response to tolbutamide. In Vascular and neurological changes in early diabetes. Eds, RA Camerini-Davalos and HS Cole. New York, Academic Press, pp 521-531. Keen, H, GA Rose, DA Pyke, D Boyns, C Chlouverakis, and S Mistry 1965. Blood sugar and arterial disease. Lancet 2:505-508. Kessler, II 1971. Mortality experience of diabetic patients A twenty-six year follow-up study. Amer J Med 51:715-724. Kipnis, DM 1968. Insulin secretion in diabetes mellitus. Ann Int Med 69:891-901. Klimt, CR, GL Knatterud, CL Meinert, and TE Prout (for the University Group Diabetes Program) 1970. A study of the effect of hypoglycemic agents on vascular complications in patients with adult onset diabetes. I. Design, methods and baseline results. Diabetes 19: (Supplement 2)747-783. Kuo, PT 1968. Current metabolic-genetic interrelationships in human atherosclerosis with therapeutic considerations. Ann Int Med 68:449-466. Kuo, PT, and DR Bassett 1965. Dietary sugar in the production of hypertriglyceridemia. Ann Int Med 62:1199-1212. Kuo, PT, and LY Feng 1970. Study of serum insulin in atherosclerotic patients with endogenous hypertriglyceridemia (Types III and IV hyperlipoproteinemia). Metabolism 19:372-380. Kwaan, HC, JA Colwell, S Cruz, N Suwanwela, and JG Dobbie 1972. Increased platelet aggregation in diabetes mellitus. J Lab Clin Med 80:236-246. Lal, HB, and AL Bahl 1967. Association of diabetes mellitus, hypertension and coronary heart disease. Ind Heart J 19:96-104. Lebovitz, HE, KT Shultz, ME Matthews, and R Scheele 1969. Acute metabolic responses to myocardial infarction. Changes in glucose utilization and secretion of insulin and growth hormone. Circ 39:171-181. Levy, RI, and CJ Glueck 1969. Hypertriglyceridenmia, diabetes mellitus and coronary vessel diseases. Arch Int Med 123:220-228. Liebow, IM, HK Hellerstein, and M Miller 1955. Arteriosclerotic heart disease in diabetes mellitus. Amer J Med 18:438-447. Maempel, JVZ 1969. ‘The etiological role of diabetes mellitus in cardiovascular disease. Israel J Med Sci 5:675-679. Major, SG 1929. Blood pressure in diabetes mellitus. Arch Int Med 44:797-812. Marks, HH 1965. Longevity and mortality of diabetics. Amer J Pub Hlth 55:416-423. Martin, FIR, and AE Stocks 1968. Insulin sensitivity and vascular disease in insulin- dependent diabetics. Brit Med J 2:81-82. Masi, AT 1972. Endocrine factors and risk of venous thrombosis. Mil Mem Fund Quart 50: (Part 2)46-59. Mayne, EE, JM Bridges, and JA Weaver 1970. Platelet adhesiveness, plasma fibrinogen and factor VIII levels in diabetes mellitus. Diabetologia 6:436-440. 107. 108. 109. 110. 111. 112. 113. 114. 115, 116. 117. 118. 119. 120. 121. 122. 123. 124. New Research Directions 209 McKiddie, MT, KD Buchanan, and IA Hunter 1969. Plasma insulin studies in two hundred patients with diabetes mellitus. Quart J Med, New Series 38:445-465. Meinert, CL, GL Knatterud, TE Prout, and CR Klimt (for the University Group Diabetes Program) 1970. A study of the effects of hypoglycemic agents on vascular complications in patients with adult-onset diabetes. II. Morgality results. Diabetes 19:(Supplement 2) 789-830. Morris, GC, Jr, GJ Reul, JF Howell, ES Crawford, DW Chapman, HL Beazley, WL Winters, PK Peterson, and JM Lewis 1972. Follow-up results of distal coronary artery bypass for ischemic heart-disease. Amer J Card 29:180-185. Mosenthal, HO, and E Barry 1950. Criteria for interpretation of normal glucose tolerance tests. Ann Int Med 33:1175-1194. Mouratoff, GJ, NV Carroll, and EM Scott 1967. Diabetes mellitus in Eskimos. J. Amer Med Assoc 199:961-966. Munro, HN, JC Eaton, and A Glen 1949. Survey of a Scottish diabetic clinic. J Clin Endocrin 9:48-78. Murphy, EA, and JF Mustard 1962. Coagulation tests and platelet economy in atherosclerotic and control subjects. Circ 25:114-125. Najenson, T, L Mendelson, H Selibiansky, R Don, and U Sandbank 1970. Diabetes and cerebrovascular accidents. Israel J Med Sci 6:598-604. New, MI, TM Roberts, EL Bierman, and GG Reader 1963. The significance of blood lipid alterations in diabetes mellitus. Diabetes 12:208-212. Nikkila, EA, and MR Taskinen 1970. Hypertriglyceridemia and insulin secretion, a complex causal relationship. In Atherosclerosis: Proceedings of the second international symposium. Ed, RJ Jones. New York, Heidelberg, Berlin, Springer-Verlag, pp 220-230. Norddy, A, and JM Rodset 1970. Platelet phospholipids and their function in patients with juvenile diabetes and maturity onset diabetes. Diabetes 19:698-702. Ostrander, LD, Jr 1970. Hyperglycemia and vascular disease in Tecumseh, Michigan. In Early diabetes. Eds, RA Camerini-Davalos and HS Cole. New York, Academic Press, pp 365- 370. Ostrander, LD, Jr, WD Block, DE Lamphiear, and FH Epstein 1973. Altered carbohydrate and lipid metabolism and coronary heart disease among men in Tecumseh, Michigan. In Vascular and neurological changes in early diabetes. Eds, RA Camerini-Davalos and HS Cole. New York, Academic Press, pp 73-81. Ostrander, LD, Jr, T Francis, Jr, NS Hayner, MO Kjelsberg, and FH Epstein- 1965. The relationship of cardiovascular disease to hyperglycemia. Ann Int Med 62:1188-1198. Ostrander, LD, Jr, DE Lamphiear, WD Block, BC Johnson, and FH Epstein (in press). Biochemical precursors of atherosclerosis among apparently healthy men in general popula- tion, Tecumseh, Michigan. Arch Int Med. Ostrander, LD, Jr, BJ Neff, WD Block, T Francis, Jr, and FH Epstein 1967. Hyperglycemia and hypertriglyceridemia among persons with coronary heart disease. Ann Int Med 67:34-41. Paasikivi, J 1973. Long-term treatment of patients with abnormal intravenous glucose tolerance after myocardial infarction. In Vascular and neurological changes in early diabetes. Eds, RA Camerini-Davalos and HS Cole. New York, Academic Press, pp 533-538. Partamian, JO, and RF Bradley 1965. Acute myocardial infarction in 258 cases of diabetes. Immediate mortality and five-year survival. New Eng J Med 273:455-461. 210 125. 126. 128. 129. 130. 131. 132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 142. 143. Diabetes Mellitus Pell, S, and CA D'Alonzo 1967. Some aspects of hypertension in diabetes mellitus. J Amer Med Assoc 202:10-16. Pell, S, and CA Alonzo 1970. Factors associated with long-term survival of diabetics. J Amer Med Assoc 214:1833-1840. Perley, M, and DM Kipnis 1966. Plasma insulin responses to glucose and tolbutamide of normal weight and obese diabetic and nondiabetic subjects. Diabetes 15:867-874. Pickering, G 1964. Pathogenesis of myocardial and cerebral infarction: Nodular arteriosclerosis. Brit Med J 1:517-529. Porte, D, Jr 1967. A receptor mechanism for the inhibition of insulin release by epinephrine in man. J Clin Invest 46:86-94. Porte, D, Jr, and EL Bierman 1969. The effect of heparin infusion on plasma triglyceride in vivo and in vitro with a method for calculating triglyceride turnover. J Lab Clin Med 73:631-648, Prout, TE 1973. Therapy of early diabetes: The University Group Diabetes Program. In Vascular and neurological changes in early diabetes. Eds, RA Camerini-Davalos and HS Cole. New York, Academic Press, pp 501-520. Pyke, DA 1968. Arterial disease in diabetes. In Clinical diabetes and its biochemical basis. Eds, WG Oakley, DA Pyke and KW Taylor. Oxford and Edinburgh, Blackwell Scientific Publications, pp 506-541. Raab, W 1960. Metabolic protection and reconditioning of the heart muscle through habitual physical exercise. Annals of Int Med 53:87-105. Raab, W 1963. Neurogenic multifocal destruction of myocardial tissue. Pathogenic mechanism and its prevention. Rev Can Biol 22:217-239. Randle, PJ, PB Garland, CN Hales, and EA Newsholme 1963. The glucose fatty-acid cycle. Its role in insulin sensitivity and the metabolic disturbances of diabetes mellitus. Lancet 1:785-789. Rathbone, RL, NG Ardlie, and CJ Schwartz 1970. Platelet aggregation and thrombus forma- tion in diabetes mellitus: an in vitro study. Pathology 2: 307-316. Reaven, G, AJ Calciano, RM Cody, C Lucas, and R Miller 1963. Carbohydrate intolerance and hyperlipemia in patients with myocardial infarction without known diabetes mellitus. J Clin Endocrin 23:1013-1023. Reaven, GM, RL Lerner, MP Stern, and JW Farquhar 1967. Role of insulin in endogenous hypertriglyceridemia. J Clin Invest 46:1756-1767. Reaven, GM, J Olefsky, and JW Farquhar 1972. Does hyperglycemia or hyperinsulinemia characterise the patient with chemical diabetes? Lancet 1:1247-1249. Reaven, GM, SW Shen, A Silvers, and JW Farquhar 1971. Is there a delay in the plasma insulin-response of patients with chemical diabetes mellitus? Diabetes 20:416-423. Rees, G, JD Bristow, EL Kremkau, GS Green, RH Herr, HE Griswold, and A Starr 1971. Influence of aortocoronary bypass surgery on left ventricular performance. New Eng J Med 284:1116-1120. Rimoin, DL 1969. Ethnic variability in glucose tolerance and insulin secretion. Arch Int Med 124:695-700. Rimoin, DL, and JH Saiki 1968. Diabetes mellitus among the Navajo. II. Plasma glucose and insulin responses. Arch Int Med 122:6-9. 144. 145. 146. 147. 148. 149. 150. 151. 152. 153. 154. 135. 156. 157. 158, 159. 160. 161. 162. New Research Directions 211 Robertson, WB, and JP Strong 1968. Atherosclerosis in persons with hypertension and diabetes mellitus. Lab Invest 18:538-551. Root, HF, EF Bland, WH Gordon, and PD White 1939. Coronary atherosclerosis in diabetes mellitus. A postmortem study. J Amer Med Assoc 113:27-30. Root, HF, and TP Sharkey 1936. Arteriosclerosis and hypertension in diabetes. Ann Int Med 9:873-882. Saiki, JH, and DL Rimoin 1968. Diabetes mellitus among the Navajo. I. Clinical features. Arch Int Med 122:1-5. Salans, LB, JL Knittle, and J. Hirsch 1968. The role of adipose cell size and adipose tissue insulin sensitivity in the carbohydrate intolerance of human obesity. Journal of Clin Invest 47:153-165. Santen, RJ, PW Willia, III, and SS Fajans 1972. Atherosclerosis in diabetes mellitus. Correlations with serum lipid levels, adiposity, and serum insulin level. Arch Int Med 130:833-843. Schor, S 1971. The University Group Diabetes Program. A statistician looks at mortality results. J Amer Med Assoc 217:1671-1675. Seltzer, HS, EW Allen, AL Herron, Jr, and MT Brennan 1967. Insulin secretion in response to glycemic stimulus: Relation of delayed insulin release to carbohydrate intolerance in mild diabetes mellitus. J Clin Invest 46:323-335. Sharkey, TP 1971. Diabetes mellitus--present problems and new research. The heart and vascular disease. J Amer Diet Assoc 58:336-344. Sharma, B, PA Majid, BC Pakrashi, JRW Dykes, and SH Taylor 1970. Insulin secretion in heart failure. Brit Med J 2:396-398. Shaw, S, GD Pegrum, S Wolff, and WL Ashton 1967. Platelet adhesiveness in diabetes mellitus. J Clin Path 20:845-847. Sloan, JM, JS Mackay, and B Sheridan 1970. Glucose tolerance and insulin response in atherosclerosis. Brit Med J 4:586-588. Sowton, E 1962. Cardiac infarction and the glucose tolerance test. Brit Med J 1:84-86. Spodick, DH 1973. The surgical mystique and the double standard. Controlled trials of medical and surgical therapy for cardiac disease: Analysis, hypothesis, proposal. Amer Heart J 85:579-533, Stamler, J, DM Berkson, and HA Lindberg 1972. Risk factors: Their role in the etiology and pathogenesis of the atherosclerotic diseases. In: The Pathogenesis of Atherosclerosis. Eds, RW Wissler, and JC Geer. Baltimore, Williams and Wilkins, pp 67-69. Stein, JH, KM West, JM Robey, DF Tirador, and GW McDonald 1965. The high prevalence of abnormal glucose tolerance in the Cherokee Indians of North Carolina. Arch Int Med 116:842-845. Stipa, S, and FC Wheelock 1971. A comparison of femoral artery grafts in diabetic and nondiabetic patients. Amer J Surg 121:223-228. Stocks, AE, and FIR Martin 1969. Insulin sensitivity and vascular disease in maturity- onset diabetics. Brit Med J 4:397-398. Stout, RW 1973. The role of insulin in the development of atherosclerosis. In Vascular and neurological changes in early diabetes. Eds, RA Camerini-Davalos and HS Cole, New York, Academic Press, pp 41-47. 212 Diabetes Mellitus 163. 164. 165. 166. 167. 168. 169. 170. 171. 172. 173. 174. 175. 176. Strandness, DE, Jr, RE Priest, and GE Gibbons 1964. Combined clinical and pathologic study of diabetic and nondiabetic peripheral arterial disease. Diabetes 13:366-372. Tzagournis, M, R Chiles, JM Ryan, and TG Skillman 1968. Interrelationships of hyper- insulinism and hypertriglyceridemia in young patients with coronary heart disease. Circ BR5 1156-1168. Tzagournis, M, JF Seidensticker, and GJ Hamwi 1967. Serum insulin, carbohydrate, and lipid abnormalities in patients with premature coronary heart disease. Ann Int Med 67:42-47. Vihert, AM, VS Zhdanov, and EE Matova 1969. Atherosclerosis of the aorta and coronary vessels of the heart in cases of various diseases. J Ather Res 9:179-192. Waddell, WR, and RA Field 1960. Carbohydrate metabolism in atherosclerosis. Metabolism 9:800-806. Wahlberg, F 1962. The intravenous glucose tolerance test. Acta Medica Scandinavica 171:1-7. Waitzkin, L 1967. Unknown diabetes mellitus among apparently healthy men with ''non- specific" T wave abnormalities--in a mental hospital. Diabetes 16:722-727. Weaver, JA, SK Bhatia, D Boyle, DR Hadden, and DAD Montgomery 1970. Cardiovascular state of newly discovered diabetic women. Brit Med J 1:783-786. Wessler, S, and NR Silberg 1953. Studies in peripheral arterial occlusive disease. IT. Clinical findings in patients with advanced arterial obliteration and gangrene. Circ 7:810-818. West, K, and JD Kalbfleisch 1971. Influence of nutritional factors on prevalence of diabetes. Diabetes 20:99-108. Wheelock, FC, Jr, and HS Filtzer 1969. Femoral grafts in diabetics. Resulting conserva- tive amputations. Arch Surg 99:776-780. Wilson, DE, PH Schreibman, VC Day, and RA Arky 1970. Hyperlipidemia in an adult diabetic population. J Chron Dis 23:501-506. Winegrad, AI, AD Morrison, and RS Clements, Jr 1973. Polyol pathway activity in aorta. In Vascular and neurological changes in early diabetes. Eds, RA Camerini-Davalos and HS Cole. New York, Academic Press, pp 117-124. } Yalow, RS, SM Glick, J Roth, and SA Berson 1965. Plasma insulin and growth hormone levels in obesity and diabetes. Ann New York Acad Sci 131:357-371. 16 OCULAR COMPLICATIONS IN DIABETES Morton E. Smith and Bernard Becker I. BACKGROUND Visual disability in diabetes can be caused by cataracts, refractive errors, and glaucoma, but the most important ocular complication which gives rise to blindness is diabetic retinopathy; a term which encompasses all the pathologic phenomena in the retina directly related to the presence of diabetes. Diabetic retinopathy is further divided into two major categories--the nonproliferative type and the proliferative type; the hallmark of the latter being the presence of newly formed blood vessels (neovascularization). Pa | A. Scope of the problem. Diabetic retinopathy is the fourth leading cause of all legal blindness (visual acuity of 20/200 or worse) in the United States today (70). Furthermore, it is on the rise as a cause of blindness because senile cataracts and glaucoma are reduced through better delivery of health care and as more diabetics live longer. The latter point is appreciated by the historical fact that before the discovery of insulin by Banting and Best in 1922, reports of proliferative dia- betic retinopathy were infrequent. In that era, diabetics did not live long enough to develop severe retinopathy or become blind from their disease. In 1930 less than 1 percent of newly reported cases of blindness in the United States were due to diabetes. Today, 15 percent of all new cases of blindness are due to diabetes; and in the 40 to 60 age group, diabetes is the most common cause of newly reported cases of blindness. Almost 40 percent of all diabetics have some degree of diabetic retinopathy, and 2 percent of all diabetics are blind (30). There are many factors which determine the occurrence of clinically detectable diabetic retinopathy, but the most important are the duration of the diabetes and the age of the patient at the time of diagnosis. If a patient is diagnosed as a diabetic at age 30, then there is a 10 percent chance he will have some degree of diabetic retinopathy by age 37, a 50 percent chance by age 45, and a 90 percent chance by age 55. This concept is slightly modified by age, i.e., if the diabetes is diagnosed before the patient is 30 and the disease lasts 5 to 9 years, the risk of diabetic retinopathy is 2 percent per year, ... but if diabetes is established after the age of 30 and the disease lasts 5 to 9 years, the risk of diabetic retinopathy is 7 percent per year. The mere existence of diabetic retinopathy does not necessarily portend a visual handicap or a poor visual prognosis. Many of these individuals have and retain excellent to adequate vision. In fact, one of the frustrations of the ophthalmologist is the inability to predict which diabetic patients will remain stable and which ones will deteriorate. Visual prognosis is Note: It is estimated from an index of prevalence of economically disadvantaged persons from diabetic retinopathy in 1968, that the total costs, which included health services, lost produc- tivity, and benefits, amounted to an absolute minimum figure of $260,000,000 per year. With rising inflation since 1968, the calculation would probably come to a 1974 minimal estimate of about $338,000,000 per year. 213 214 Diabetes Mellitus estimated not only on the duration of the disease but also on what the visual acuity is at the time of diagnosis and whether or not the retinopathy is nonproliferative or proliferative (55). From the statistics available, a general picture can be drawn. If, at the time of diagnosis of diabetes, the patient has minimal retinopathy of the nonproliferative type and the vision is good (20/40 or better), then the risk of blindness in 5 years is less than 10 percent. The risk of blindness in 5 years is greater if the vision is already bad at the time of diagnosis (20/80 - 20/200), and especially if the retinopathy is of the proliferative type; this risk being about 50 percent. It has also been pointed out that if one eye is already blind from diabetic retinopathy, there is a 50 percent chance the other eye will have a significant loss of vision within 1 year (56). The magnitude of the problem is further emphasized when the prevalence of blindness in the diabetic is compared to the prevalence of blindness in the general population. The prevalence of blindness from all causes at age 50 is 0.15 percent; but in any diabetic at age 50 who has had the disease for over 15 years, the prevalence of blindness is 3.5 percent, i.e., the dia- betic is 23 times more likely to be blind than his nondiabetic counterpart (30). The presence of severe diabetic retinopathy also bears on the prognosis for life. Berkow et al (5) made an interesting observation from reviewing records of patients who sought a guide dog at Seeing Eye, Incorporated, where it had been noted that blind diabetics rarely returned for a second dog. They found that there were 85 blind diabetics who had died since getting their first guide dog, and that the average life span after the onset of blindness was only 6 years. In summary, diabetes is the most important systemic disease giving rise to blindness. The visual prognosis in a diabetic is influenced by the age of the patient and the vision at the time of diagnosis, and more importantly by the duration of the disease and the type of retinopathy, i.e., proliferative versus nonproliferative. Blindness due to diabetic retinopathy is an ominous finding in regard to the prognosis for life. One out of every 40 diabetics is blind, and two out of every 1,000 become blind each year. This represents over 10 times the risk of blindness from all causes in the general population. B. Characteristic appearance and course of diabetic retinopathy. 1. Nonproliferative retinopathy (Figs. 1 and 2). a. Pathology of the capillary bed. Nonproliferative diabetic retinopathy (also commonly referred to as background retinopathy) is first manifested by the appearance of the microaneurysm. These globular or fusiform out- pouchings occur from one side of the capillary wall, range from 10-200 microns, and vary in color from deep purple-red to yellow-white. Although microaneurysms can be seen by routine ophthalmoscopy, fluorescein angiography’ has greatly facilitated their recognition and has shown that they are particularly prominent around the edge of areas of nonperfused, obliterated capillaries. It is here in the capillary bed that the initial changes are believed to occur. The 'shunt theory" (13) suggests that first there is loss of capillary pericytes,’ resulting in loss of 1A technique whereby fluorescein dye is injected into an arm vein and photgraphs taken when the dye reaches the retinal vessels. 23pecial cells (also referred to as mural cells) in the walls of capillaries. Ocular Complications 215 capillary tone, microaneurysm formation, and dilation of some capillaries which shunt or "steal" blood away from other capillaries, leading to areas of obliteration of adjacent vascular chan- nels. Other investigators (18,3) feel that the sequence is reversed, i.e., capillary oblitera- tion followed then by secondary dilation of vascular channels. Ashton postulated that the capil- laries may actually be squeezed shut by the surrounding edematous retinal tissue, or that arterio- lar perfusion pressure is altered as the result of degenerative changes in the arteriolar wall. : FIGURE 2. Fluorescein angiogram shows scattered micro- FIGURE 1. Ophthalmoscopic view showing aneurysms which "fluoresce" (arrows). An area of fluor- hemorrhages (H), exudates (E), and micro- escein leakage is in the upper left of photo (L). The aneurysms (M). vein (V) shows 'beading." Regardless of the precise initial sequence of events, retinal capillary closure certainly occurs early in the natural history of diabetic retinopathy, thereby inducing tissue hypoxia which can be demonstrated by clinical and pathologic techniques. Aside from the microaneurysm, some of the other clinical manifestations include the following changes. b. Edema, exudates, hemorrhages, arteriolar and venous changes. Retinal edema is common in both nonproliferative and proliferative retinopathy. It often involves the macula’ and is the most common cause of reduced vision in nonproliferative retinopathy. Clinically, the macula appears swollen and angiographs show diffuse intraretinal leakage of fluorescein; presumably due to altered permeability of surrounding capillaries. "Hard, waxy exudates" is the term used to describe the pockets of extravasated protein and lipid which form in the deep layers of the retina, probably also as a result of altered capil- lary permeability. These discrete yellow-white flecks usually occur near areas of abnormal capillaries and microaneurysms, often becoming confluent to form a partial or complete circle around an edematous macula (''circinate retinopathy"). "Cotton wool spots,'" a lesion considered to be characteristic of hypertensive retinopathy, has recently been recognized as an early component of diabetic retinopathy without vascular hypertension (42). These grey-white lesions with indistinct borders result from microinfarc- tions in the superficial retinal layers; further support that closure of capillaries and/or terminal arterioles is a fundamental feature of this disorder. 3The region of the retina responsible for central vision. 216 Diabetes Mellitus Retinal hemorrhages in diabetic retinopathy can be flame-shaped in the superficial layers or the more common round or "blot'" hemorrhage in the deeper retinal layers. The arterioles may show changes which are usually the result of generalized cardiovascular disease which may accompany the diabetes. Changes in the retinal veins are also considered an early sign of dia- betic retinopathy and consist mainly of increased tortuosity and focal variation in caliber, referred to as 'beading.'" These venous changes are often found in juvenile diabetics. 2. Proliferative diabetic retinopathy (6 percent of all cases of diabetic retinopathy or 2 percent of all diabetics) Figs. 3-6. FIGURE 3. A tuft of new vessels (arrows) appears on the inner surface of the retina. (From Okun, E, et al. 1971) Reprinted with FIGURE 4. A network of neovascularization arises from permission of C. V. Mosby, St. Louis. the optic disk. The hallmark of proliferative diabetic retinopathy is the appearance of neovascularization, which drastically changes the clinical picture as well as the prognosis. It is first appre- ciated ophthalmoscopically as a 'brush'" of fine capillaries in an area where the normal capil- lary bed has been damaged or destroyed. The site of predilection for neovasculatization are the optic disk, along the course of the main vessels, and in the equatorial region. These new vessels form networks along the inner surface of the retina and follow a cycle of proliferation and regression. They evolve in three stages: 1) naked stage, in which these fine-walled vessels appear to have no connective tissue. After an interval of 1 to 4 years, they progress to 2) fibrous stage, in which there is an increase in the size of the vessels with formation of connective tissue around them. This progresses in 1 to 2 years to 3) regression and scarring when the connective tissue forms an avascular mass over the retina and disk, and the normal retinal arterioles become attenuated. Adhesions form between this network of fibrovascular proliferation and the vitreous body. When the vitreous contracts, these fragile vessels are pulled forward often resulting in vitreous hemorrhage and/or retinal detachment, the two major causes of blindness. Ocular Complications 217 FIGURE 6. A dense connective tissue mass with neovascularization obscures retinal details. (From Okun, E, et al. 1971) Re- FIGURE 5. A large fresh hemorrhage into the vitreous printed with permission of C. V. Mosby, is seen at the upper right corner of the photo. St. Louis. The rate of progression of proliferative retinopathy is variable. Occasionally it will go into a remission or "burned out' phase in which the vitreous contraction and the recurrent hemorrhages stop, and there is an overall ischemia of the retina. In summary, capillary closure and tissue hypoxia occur early in diabetic retinopathy, fol- lowed by or coincident with edema, hemorrhages, exudates and eventually, in some cases, neovas- cularization. This latter ominous event may eventually go on to vitreous hemorrhage and retinal detachment with resultant blindness. C. Therapy of diabetic retinopathy (53). In order to evaluate the efficacy of various forms of therapy, it was necessary to estab- lish a universally accepted classification of diabetic retinopathy. The most recent one, known as the Airlie House classification, is an attempt to present a framework for prospective studies with the hope that objective comparisons can be made. This classification is based on the pho- tographic and ophthalmoscopic findings just described above (17). 1. Diabetic control. Probably the most perplexing and frustrating question concerns the relationship between medical control of the diabetes and the occurrence and progression of diabetic retinopathy. ° Meaningful interpretation of data is hindered by the lack of criteria for what constitutes ''good control." There is some statistical evidence which implies that better control of diabetes in patients under age 60 reduces the frequency or delays the appearance of retinopathy by a few years. Once retinopathy is established, however, control appears to have little or no effect (41). 218 Diabetes Mellitus 2. Pituitary ablation (29). In 1953, Poulsen reported a dramatic improvement in the retinopathy of a young diabetic woman after she developed Sheehan's syndrome” following the birth of a stillborn. Deliberate ablation of the pituitary by various means followed, and varying degrees of success have been reported (44). Assessment of these results, however, is difficult in light of the occasional spontaneous improvement of diabetic retinopathy which, of course, clouds the issue in evaluating any mode of therapy. Selection of patients for this treatment is quite strict. These diabetics should be highly motivated individuals capable of following a demanding regimen of replacement therapy. Besides having a rapidly advancing proliferative retinopathy, but with macular function preserved in at least one eye, there must be absence of advanced renal, cardiovascular, and neurological dis- ease. } The possibility of improvement must be balanced against the risks and disadvantages of the procedure, and with the data available, pituitary ablation cannot be considered the procedure of choice at this time except perhaps for a small number of carefully selected patients. 3. Photocoagulation (29). Photocoagulation is the use of high intensity light from either a xenon arc or a laser to produce a "burn" with subsequent scarring of the tissues in the posterior portion of the eye. The method is thought to be beneficial in diabetic retinopathy is several ways: 1) areas of critical ischemia are destroyed, thus decreasing the stimulus to neovascularization; 2) new vessels are destroyed, thus removing the threat of hemorrhage; 3) leakage from permeable capil- laries is lessened; 4) chorioretinal adhesions limit the ease of retinal detachment; 5) meta- bolic requirements of the retina may be diminished. Photocoagulation is used primarily in proliferative retinopathy, but recently has been shown to be beneficial in nonproliferative retinopathy when the vision is impaired due to macular edema (60). In the latter instance, photocoagulation can aid by sealing leaks demon- strated by fluorescein angiography. Results have been encouraging and most workers agree with the report of Okum (52) where in a small series of patients in which both eyes had symetrical involvement by proliferative retinopathy but only one eye was treated, almost half of the treated eyes improved while less than 10 percent of the untreated eyes improved. Long-term, randomized controlled studies are still needed, and an important step in this direction has been initiated by the Diabetic Reti- opathy Study under the aegis of 16 medical centers in cooperation with the National Eye Insti- tute (45). 4. Other modes of therapy. Many other modes of medical therapy have been used against diabetic retinopathy, but be- cause long-term controlled studies have been conspicuously lacking, these methods cannot be evaluated in any meaningful way and only brief mention is made here. These include rutin, testosterone, heparin, phenindione, vitamin B;,, para-aminosalicylic acid, corn oil, low fat diet, and clofibrate. It appears that the last three mentioned approaches do reduce the amount of exudation, but this is not accompanied by any visual improvement. Calcium dobesilate, a drug being used in Europe, appears to show promise in retarding the progression of nonproliferative retinopathy by restoring normal capillary function (61). “A syndrome in which the pituitary gland is destroyed. Ocular Complications 219 In summary, the therapy of diabetic retinopathy is less than satisfactory at this time. No proven medical therapy exists, and the role of 'good control" remains an enigma. Pituitary ablation does not appear to have proven its worth in light of the risks involved. Photocoagulation appears to be the most promising approach to palliation of proliferative retinopathy, but its true value is yet to be determined. D. Cataracts, glaucoma. 1. Cataracts. The association between cataracts and diabetes is not clearcut since it is not al- ways possible to attrubute the presence of a cataract in an adult to his diabetes. It is cus- tomary and fair, however, to assume that cataracts occurring in diabetics under the age of 40 are presumably related to the systemic disease. Statistical evidence also exists to support the notion that cataracts in older individuals occur more frequently in diabetics than in nondiabet- ics, and that cataracts in diabetics mature more rapidly than cataracts in nondiabetics (8,51). Although the cataracts can take any nonspecific form, the so-called '"snowflake' pattern of opacity is considered characteristic in diabetes and may be the first manifestation of the dis- ease. Many factors are probably involved but sorbitol® accumulation in the lens appears to play an important role in these diabetic cataracts (40,24,68). The osmotic effects of sorbitol accumulation in the lens may also be responsible for the transient myopia which so often occurs in diabetics. Therapy for the cataract consists of surgical removal and does not differ from conventional cataract removal. : 2. Glaucoma. a. Secondary glaucoma associated with neovascularization of the iris. Rubeosis iridis is the clinical term used to describe new vessel formation on the anterior surface of the iris. This enigmatic phenomenon occurs in a variety of ocular diseases including diabetes where it is almost invariably associated with the presence of severe proliferative retinopathy. Estimates of the frequency of rubeosis iridis in diabetics vary widely, but a figure of 1 percent in all diabetics appears reasonable (49). This fine network of new blood vessels can be seen clinically around the pupillary margin, over the surface of the iris, and in the anterior chamber angle (the anatomical angle formed by the peripheral iris and peripheral cornea). Eventually, this network of blood vessels and its accompanying fibrous tissue contracts and pulls the peripheral iris up over the outflow channels of the peripheral cornea to form dense adhesions, subsequent obstruction to the outflow of aqueous humor, and an intractable glaucoma (37). The treatment of this type of glaucoma is un- rewarding and most of these already blind eyes need to be surgically removed because of unre- lenting pain. b. Open angle glaucoma and diabetes. Primary open angle glaucoma (''chronic simple glaucoma') is the most common form of glaucoma that exists in the general population and is believed to be due to as yet unexplained degenera- tive processes occurring in the channels of the eye which allow for egress of aqueous humor. Recent clinical studies (4) have shown the following statistical relationships to exist between primary open angle glaucoma and diabetes: 1) primary open angle glaucoma is more prevalent in diabetics than in nondiabetics; 2) proliferative retinopathy occurs less often in diabetic patients with primary open angle glaucoma than in nonglaucomatous diabetics; 3) diabetes 5Sugar alcohol analogue of glucose. See Chapter 14. 220 Diabetes Mellitus and positive glucose tolerance tests are more prevalent in the glaucoma population; 4) the glaucoma population with a positive glucose tolerance test appears to be more susceptible to glaucomatous visual field loss than those with a negative glucose tolerance test; 5) positive glucose tolerance tests are more prevalent in glaucomatous patients with lower intraocular pressures. IT. EXPERIMENTALLY INDUCED RETINOPATHIES. Although diabetes can be induced in animals in a variety of ways, complete pathologic changes characteristic of human diabetic retinopathy seldom develop and most of the changes are minimal. This is also true of spontaneous diabetes in animals. Table 1 summarizes these find- ings. Engerman (23) has shown that diabetic dogs kept in "good control' developed less retinop- athy over a period of years compared to dogs purposely kept in "poor control." ITI. CURRENT STUDIES, GAPS IN OUR KNOWLEDGE, AND THOUGHTS FOR THE FUTURE. The exact initial sequence of events leading to diabetic retinopathy still remains a mystery. If, as Cogan and Kuwabara (13) have suggested, the initial event is the loss of the capillary pericyte, then what causes this selective "drop-out" of these cells in diabetes? The exact function of this particular cell still needs to be determined. Although a vasomotor func- tion has been attributed to this cell, evidence is lacking to support this contention. Recent studies have shown that this cell may have a phagocytic function (19). If capillary obliteration is the initial event, and if surrounding edematous retinal tissue is responsible for this capillary shutdown (3), then it becomes interesting to speculate on pos- sible causes for this retinal edema. We have seen how the sorbitol pathway is responsible for fluid accumulation in the lens. Could the same process be occurring in the retinal tissues? The need for further biochemical studies on the diabetic retina becomes obvious. The nature of the microaneurysm needs further elucidation. Although most diabetic micro- aneurysms are thin walled, some have a thick wall giving an intense staining reaction for glyco- proteins. Endothelial cell degeneration and proliferation, plus basement membrane thickening with vacuolization, are also part of the microscopic picture. The finding that fluorescent in- sulin binds specifically to these microaneurysms suggests the possibility of an autoimmune re- action, but this may be secondary rather than causal (15,46). Ashton has suggested that the microaneurysm may represent an abortive attempt at neovascularization. The stimulus for neovascularization is still unknown, but it has been suggested that a "vasoformative factor" may be liberated from hypoxic foci in the retina (2). Such a factor may also be responsible for the development of rubeosis iridis. What role does pressure play; both the pressure within the vascular system as well as intra- ocular pressure? It has already been pointed out that proliferative retinopathy occurs less often in diabetics with primary open angle glaucoma than in nonglaucomatous diabetics. Also, diabetics with no retinopathy tend to have a lower systemic blood pressure than diabetics with retinopathy. Furthermore, when the retinopathy is asymetrical in both eyes of a patient, the less affected eye is usually associated with a decreased retinal arterial diastolic pressure (26). Duane (21) correlated the progression of diabetic retinopathy with the ratio of the intra- retinal arteriolar pressure to the intraocular pressure. Conditions which increased the numer- ator (intra-arteriolar pressure) such as systemic hypertension, or decreased the denominator TABLE 1. Experimental Diabetic Retinopathy* Animal Monkey Dog Dog Dog Dog Dog Chinese hamster Rat Rat Rat Rat Rat Carp Carp Carp Mode of Production Alloxan Spontaneous Spontaneous Spontaneous Alloxan Growth hormone Growth hormone Pancreatectomy Alloxan Alloxan and iminidiproprionitrile Cortisone and growth hormone Streptozotocin Spontaneous (Sekoke disease) Alloxan Hydrocortisone Lesions Microaneurysms Microaneurysms, exudates, loss of pericytes Microaneurysms Microaneurysms Microaneurysms, pericyte loss, acellular capillaries Microaneurysms, pericyte loss, acellular capillaries Microaneurysms Microaneurysms Intravitreal new vessels Acellular capillaries Microaneurysms and acellular capillaries Microaneurysms, basement membrane changes Dilation of vessels Dilation of vessels Dilation of vessels Authors Gibbs et al (1966) Patz and Maumenee (1962) Gepts and Toussaint (1967) Sibay and Hausler (1967) Engerman and Bloodworth (1965) Engerman and Bloodworth (1965) Hausler et al (1963) Musacchio et al (1964) Toussaint (1966) Heath and Rutter (1966) Agrawal et al (1966) Leuenberger et al (1971) Yokote (1970) Yokote (1970) Yokote (1970) *Partially adapted from Caird, F. I., et al, 1969, Diabetes and the Eye, Table 4.1, p. 54. Tzz suorzeorTdwo) Iernop 222 Diabetes Mellitus (intraocular pressure) as in postoperative hypotony, tended to aggravate the retinopathy. Although it is unlikely that changes in intraocular pressure or systemic perfusion pressure are causative, it is probable that these alterations influence the point at which vascular insuffi- ciency and tissue damage occur. Direct measurement of blood flow in retinal arterioles holds promise in answering some of these questions (36). Much recent attention has been directed toward the possible role of clotting factors, particularly intravascular platelet aggregation (34). Recent evidence implies that there is an "enhancing factor" in the plasma of diabetics with retinopathy which is responsible for in- creased intravascular aggregation of platelets which in turn may produce the retinal capillary obstruction. This "enhancing factor" is not present in diabetics without retinopathy or in non- diabetics. On the basis of these findings, preliminary clinical trials using aspirin (a known inhibitor of platelet aggregation) are being conducted on diabetics (11,20). Other metabolic parameters which have been implicated as being altered in diabetics with retinopathy but not in diabetics without retinopathy or in nondiabetics, and which need confir- mation and/or refinement include: erythrocyte aggregation, fibrinogen concentration, serum proteins (67), serum lipids (58), growth hormone (62), and serum prostaglandins (69). A better animal model of diabetic retinopathy is urgently needed and progress will be made if conditions can be found which, when superimposed upon experimental diabetes, accelerate the rate of pathological change in the retina (33). From the clinical point of view, it has been pointed out that photocoagulation of prolif- erative retinopathy is only palliative and has no effect on the basic disease process. Although improved instrumentation may improve the efficacy of photocoagulation, it is important to con- tinue the search for better medical therapy which could retard the progression of diabetic retinopathy before the onset of the proliferative state. Reference has already been made to the use of calcium dobesilate for restoration of normal capillary function and inhibitors of platelet aggregation such as aspirin. If better medical therapy can be found, or if the role of ''good control' becomes better understood, it becomes important to recognize the very earliest clinical signs or symptoms of diabetic retinopathy, perhaps even prior to ophthalmoscopically apparent disease. Encouraging studies from this point of view include such findings as impairment of color vision (39), electro-oculography changes (35), electroretinography changes (64), decreased flicker fusion response (12), and the identification of small blind spots (scotometry) (59). In summary, attention centers around: 1) the initial events leading to capillary obliter- ation and the exact nature of the microaneurysm; 2) what role do other parameters play, e.g., intraocular pressure, vascular perfusion pressure, intravascular clotting factors, serum con- stituents; 3) production of a better experimental mode; and 4) detection of the earliest func- tional changes prior to morphologic changes. CONCLUSION From the foregoing discussion it becomes obvious that any successful approach to the pre- vention and therapy of ocular complications of diabetes will come only with a better understand- ing of the initial events which occur in the diabetic retina and other ocular tissues. The ophthalmologist must remain an important member of the interdisciplinary team, working closely Ocular Complications 223 with the biochemist and pathologist as well as acting as a liaison between basic science and clinical application. Pragmatically speaking, this role can best be accomplished if financial resources are made available to enable interested research ophthalmologists to become special- ists just in ocular complications of diabetes. REFERENCES 1. Agrawal, PK, LP Agarwal, and HD Tandon 1966. Experimental diabetic retinopathy in albino rats. Orient Arch Ophth 4:68. 2. Ashton, N 1961. Neovascularization in ocular disease. Tr Ophth Soc, UK 81:145. 3. Ashton, N 1963. Studies of the retinal capillaries in relation to diabetic and other retinopathies. Brit J Ophth 47:521-538. 4. Becker, B 1971. Diabetes mellitus and primary open angle glaucoma. Am J Ophth 71:1-16. 5. Berkow, J, R Shugarman, AE Maumenee, and A Patz 1965. A retrospective study of blind dia- betic patients. JAMA 193:870. 6. Bloodworth, JMB 1962. Diabetic retinopathy. Diabetes 11:1-22. 7. Brotherman, DP 1972. Diabetic retinopathy: A perspective. Survey Ophth 16:359-370. 8. Burditt, AF, and FI Caird 1968. The natural history of lens opacities in diabetes. Brit J Ophth 52:433. 9. Caird, FI, A Pirie, and TG Ramsell 1969. Diabetes and the eye. Oxford and Edinburgh, Blackwell Scientific Publishers. 10. Carroll, WW 1972. Acuity variations in diabetes mellitus. Med Col Virginia Quart 8:286- 288. 11. Carroll, WW, and W Geeraets 1972. Diabetic retinopathy and salicylates. Ann Ophth 72: 1019-1046. 12. Chochinov, RH, LE Ullyot, and J Moorhouse 1972. Sensory perception thresholds in patients with juvenile diabetes and their close relatives. N Eng J Med 286:1233-1237. 13. Cogan, DG, and T Kuwabara 1963. Capillary shunts in the pathogenesis of diabetic reti- nopathy. Diabetes 12:293-300. 14. Cogan, DG 1964. Current concepts: Diabetic retinopathy. N Eng J Med 270:787-788. 15. Coleman, SL, B Becker, S Canaan, and L Rosenbaum 1962. Fluorescent insulin staining of the diabetic eye. Diabetes 11:375-377. 16. Cunha-Vaz, JG 1973. Diabetic retinopathy. Human and experimental studies. Tr Ophth Soc UK 92:111-124. 17. Davis, MD, E Norton, and FL Myers 1968a. The Airlie classification of diabetic retinopathy. In Symposium on the Treatment of Diabetic Retinopathy, edited by Morton Goldberg and Stuart Fine. Public Health Service Publication No. 1890, Washington, D.C., U.S. Government Printing Office, pp 7-22. 18. Davis, M, F Myers, R Engerman, G de Venecia, and Y Magli 1968b. Clinical observations con- cerning the pathogenesis of diabetic retinopathy. In Symposium on the Treatment of Dia- betic Retinopathy, edited by Morton Goldberg and Stuart Fine. Public Health Service Publication No. 1890, Washington, D.C., U.S. Government Printing Office, pp 47-53. 224 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. Diabetes Mellitus De Oliviera, LFN 1966. Pericytes in diabetic retinopathy. Brit J Ophth 50:134. Dobbie, JG, H Kwaan, J Colwell, and N Suvanwela 1972. Unknown blood factor appears to play role in diabetic retinopathy. JAMA 222:1356. Duane, T, T Behrendt, and R Field 1968. Net vascular pressure ratios in diabetic reti- nopathy. IN Symposium on the Treatment of Diabetic Retinopathy, edited by Morton Goldberg and Stuart Fine. Public Health Service Publication No. 1890, Washington, D.C., U.S. Government Printing Office, pp 657-663. Engerman, RL, and JMB Bloodworth, Jr 1965. Experimental diabetic retinopathy in dogs. Arch Ophth 73:205. Engerman, R 1973. Unpublished data presented as an abstract to annual meeting of Assoc Res Vision and Ophthalmology. Gabbay, KH 1973. The sorbitol pathway and the complications of diabetes. N Eng J Med 288:831-836. Garner, A 1970. Pathology of diabetic retinopathy. Brit Med Bull 26:137-142. Gay, AJ, and L Rosenbaum 1966. Retinal artery pressure in asymmetric diabetic retinopathy. Arch Ophth 75:758. Gepts, W, and D Toussaint 1967. Spontaneous diabetes in dogs and cats: A pathologic study. Diabetologia 3:249. Gibbs, GE, RB Wilson, and H Gifford 1966. Glomerulosclerosis in the long-term alloxan- diabetic monkey. Diabetes 15:258. Goldberg, MF, and SL Fine 1968. Symposium on the Treatment of Diabetic Retinopathy. Public Health Service Publication No. 1890, Washington, D.C., U.S. Government Printing Office. Goldberg, MF 1972. Natural history of diabetic retinopathy. Israel J Med Sci 8:1311-1315. Hausler, HR, TM Sibay, and B Stachowska 1963. Observations of retinal microaneurysms in a metahypophyseal diabetic Chines hamster. Am J Ophth 56:242. Heath, H, and AC Rutter 1966. Retinal angiopathy in the iminodiproprionitrile-treated alloxan-diabetic rat. Brit J Exp Path 47:116. Heath, H 1970. Experimentally induced retinopathies in relation to the problem of dia- betes. Brit Med Bull 26:151-155. Heath, H, WD Brigden, JV Canever, J Pollock, PR Hunter, J Kelsey, and A Bloom 1971. Platelet adhesiveness and aggregation in relation to diabetic retinopathy. Diabetologia 7:308-315. Henkes, HE, and AJ Houtsmuller 1965. Fundus diabeticus: An evaluation of the prereti- nopathic stage. Am J Ophth 60:662. Hickam, JB, and R Frayser 1965. A photographic method for measuring the mean retinal circulation time using fluorescein. Invest Ophth 4:876. Hohl, RD, and DM Barnett 1970. Diabetic hemorrhagic glaucoma. Diabetes 19:944-947. Kimura S, and W Caygill 1967. Vascular Complications of Diabetes Mellitus. St. Louis, CV Mosby. Kinnear, PR, PA Aspinall, and R Lakowski 1973. The diabetic eye and colour vision. Tr Ophth Soc UK 92:69-78. Kinoshita, JH 1965. Cataracts in galactosemia. Invest Ophth 4:786-799. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. Ocular Complications 225 Knowles, H. 1968. Summary of papers on relationship of retinopathy to metabolic control. In Symposium on the Treatment of Diabetic Retinopathy, edited by Morton Goldbert and Stuart Fine. Public Health Service Publication No. 1890, Washington, D.C., U.S.Government Printing Office, pp 129-131. Kohner, E, CT Dollery, and CJ Bulpitt 1969. Cotton-wool spots in diabetic retinopathy. Diabetes 16:691-704. Kohner, EM, and CT Dollery 1970a. The rate of formation and disappearance of microaneurysms in diabetic retinopathy. Tr Ophth Soc UK 90:369-374. Kohner, E, CT Dollery, TR Frazer, and CJ Bulpitt 1970b. The effect of pituitary ablation on diabetic retinopathy studied by fluorescein angiography. Diabetes 19:703-704. Kupfer, C 1973. Evaluation of the treatment of diabetic retinopathy. A research project. The Sight Saving Review, Spring, pp 17-28. Larsen, H, and AU Werner 1968. Fluorescence microscopic, autoradiographic, and histo- chemical studies on human diabetic and nondiabetic eyes. In Symposium on the Treatment of Diabetic Retinopathy, edited by Morton Goldberg and Stuart Fine. Public Health Service Publication No. 1890, Washington, D.C., U.S. Government Printing Office, pp 673-680. L'Esperance, FA 1973. Argon laser photocoagulation of diabetic retinal neovascularization (a five-year appraisal). Tr Am Acad Ophth Otol 77:6-24. Leuenberger, P, D Cameron, W Stauffacher, AE Renold, and J Babel 1971. Ocular lesions in rats rendered chronically diabetic with streptozotocin. Ophth Res 2:189-204. Madsen, P H 1971. Rubeosis of the iris and haemorrhagic glaucoma in patients with prolif- erative diabetic retinopathy. Brit J Ophth 55:368-371. Musacchio, ITL, N Palermo, and RR Rodriguez 1964. Microaneurysms in the retina of dia- betic rats. Lancet 1:146. Norm, MS 1967. Diabetes mellitus and cataracts senilis. Acta Opth (Kbh) 45:322. Okun, E 1968. The effectiveness of photocoagulation in the therapy of proliferative dia- betic retinopathy (a controlled study in 50 patients). Tr Am Acad Ophth Otol 72:246-252. Okun, E, GP Johnston, and I Boniuk 1971. Management of Diabetic Retinopathy. St. Louis, CV Mosby. Patz, A, and AE Maumenee 1962. Studies in diabetic retinopathy. I. Retinopathy in a dog with spontaneous diabetes mellitus. Am J Ophth 54:532. Patz, A 1968. Visual and systemic prognosis in diabetic retinopathy. Tr Am Acad Ophth Otol 72:253-258. Patz, A, and J Berkow 1968. Visual prognosis in advanced diabetic retinopathy. In Symposium on the Treatment of Diabetic Retinopathy, edited by Morton Goldberg and Stuart Fine. Public Health Service Publication No. 1890, Washington, D.C., U.S. Government Printing Office, pp. 87-91. Poulsen, JE 1953. Recovery from retinopathy in a case of diabetes with Simmond's disease. Diabetes 2:7. Richards, R 1962. Lipids in diabetic retinopathy. J Am Geriat Soc 10:831-842. Roth, JA 1968. New prototype central field scotometer. Brit J Ophth 52:400. Rubenstein, K, and V Myska 1972. Treatment of diabetic maculopathy. Brit J Ophth 56:1-5. Sévin, R, and JF Cuendet 1971. The action of calcium dobesilate on capillary permeability in diabetes. Ophthalmologica 162:33-40. 226 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72, 73. 74. Diabetes Mellitus Sévin, R 1972. The correlation between human growth hormone concentration in blood plasma and the evolution of diabetic retinopathy. Ophthalmologica 165:71-77. Sibay, TM, and HR Hausler 1967. Eye findings in two spontaneously diabetic related dogs. Am J Ophth 63:289. Simonsen, SE 1968. ERG in juvenile diabetics: A prognostic study. In Symposium on the Treatment of Diabetic Retinopathy, edited by Morton Goldberg and Stuart Fine. Public Health Service Publication No. 1890, Washington, D.C., U.S. Government Printing Office, pp. 681-689. Skanse, B 1963. The serum proteins in diabetes mellitus in relation to vascular complica- tions. Acta Endocrinol 43:3. Toussaint, D 1966. Lesions rétiniénnes au cours de diabéte alloxanique chez le rat. Bull Soc belg Ophtal 143:648. Van Haeringen, NJ, JA Oosterhuis, J Terpstra, and E Glasius 1973. Erythrocyte aggregation in relation to diabetic retinopathy. Diabetologiz 9:20-24. Van Heyningen, R 1959. Formation of polyols by the lens of the rat with ''sugar' cataract. Nature (London) 184:194-195. Waitzman, MB 1973. Prostaglandins and diabetic retinopathy. Exp Eye Res 16:303-307. Welch, RJ 1969. Causes of blindness in addition to MRA Registers for 1967. Proceedings of the 1969 Conference of Model Reporting Area for Blindness, Washington, D.C., U.S. Government Printing Office. Yanoff, M 1966. Diabetic retinopathy. N Eng J Med 274:1344. Yokote, M 1970. Sekoke disease, spontaneous diabetes in carp. I. Pathologic study. Bull Freshwater Fish Res Lab (Tokyo) 20:39. Tokote, M 1970. Sekoke disease, spontaneous diabetes in carp. IV. Alloxan diabetes in carp. Bull Freshwater Fish Res Lab (Tokyo) 20:147. Yokote, M 1970. Sekoke disease, spontaneous diabetes in carp. V. Hydrocortisone diabetes in carp. Bull Freshwater Fish Res Lab (Tokyo) 20:161. 1/7 RENAL DISEASE IN DIABETES MELLITUS K. Lundbaek and R. @sterby Diabetic renal disease occupies a central position in the general field of diabetic angiopathy. In the natural history of diabetes mellitus as it appears under current standard treatment, characteristic changes develop in the kidney, leading eventually to partial or complete loss of function. This development can be followed accurately with modern morphometric tech- niques and function tests. Recently biochemical methods have been added to the armamentarium of kidney research in diabetes. In the present chapter a short presentation of the problems of renal abnormalities in dia- betic patients will be given, emphasizing the important gaps in present information and under- standing, and pointing out fields of research that should, in the opinion of the authors, be promoted. A detailed discussion of the literature of the last decade dealing with renal disease in diabetes mellitus has been published recently (39). DEFINITIONS AND ALTERNATIVES IN THE STUDY OF DIABETIC RENAL DISEASE "Diabetic nephropathy" is and should be used as a clinical term to denote a chronic kidney disease with proteinuria and/or reduced renal function found in patients who have had diabetes mellitus for many years and supposed to be due to diabetic glomerulosclerosis, with or without chronic infection of the renal tissue. The clinical diagnosis of diabetic nephropathy is always tentative and based on the assumed exclusion of other causes of renal disease. There are no specific clinical features. Late in the course of disease, an otherwise uncommon picture consisting of gross proteinuria, edema, and hypertension occurs, but only in a minority of cases. "Diabetic glomerulosclerosis' and "diabetic glomerulopathy' are synonymous terms indicating the specific morphologic abnormality of the renal glomerulus. The first description of its nodular form was given by Kimmelstiel and Wilson in 1936. The diagnosis is made by microscopic study of kidney tissue, i.e., post mortem or on biopsy specimens. Renal function is abnormal in diabetic patients, but this fact cannot be used as basis for a definition of diabetic renal disease. When functional tests are performed in patients with short-term diabetes, glomerular filtration rate and other functional parameters are seen to be increased. In late stages there is a generalized decrease of renal functions. Both patterns are quite nonspecific. Recently a number of important papers have been published about the biochemical composition of the normal and the diabetic glomerular basement membrane and of the enzyme activities respon- sible for the synthesis of basement membrane glycoproteins. At the moment this field is rich in controversial problems. In human beings, biochemical anomalies can be studied only on autopsy specimens. Future technical developments may make it possible to introduce a biochemical 227 228 Diabetes Mellitus definition of diabetic renal disease. Today the clinical definition--diabetic nephropathy-- and the morphological definition-- diabetic glomerulosclerosis or glomerulopathy--are the only alternatives for the study of the natural history of diabetic renal disease, its intrinsic mechanisms, its relation to other dia- betic phenomena, and its modification under various therapeutic or prophylactic conditions. For research purposes the vague clinical term "diabetic nephropathy' can be sharpened by appropriate criteria for classification and made useful in some types of long-range, large- scale population studies. However, diabetic glomerulopathy with its precise definition and practically speaking com- plete specificity is the only acceptable object for most types of detailed studies aiming at clarification of the pathogenesis of diabetic renal disease and the possible effect of treatment or prophylaxis. It should be remembered, however, that the choice of diabetic glomerulopathy as a defini- tion for research work in human subjects entails the use of kidney biopsy. This problem is a very general one and shall not be discussed in detail here. Both percutaneous and open kidney biopsy have a certain risk, albeit a very low one. In the hands of an experienced investigator and with the proper precautions, it can be a perfectly justified procedure in important and well-designed research projects. DIABETIC NEPHROPATHY The impact of renal disease per se as a cause of suffering, invalidity, and death seems obvious, but is not too well documented. It is not prominent in patients with short or moder- ately long duration of diabetes. Long-term diabetic patients may suffer for years from tired- ness, some of them may have periods of low-grade fever and lumbar pain, but the incidence of such symptoms is hard to estimate. Once uremia sets in, the usual symptoms and signs appear, "and finally complete renal insufficiency occurs. The prevalence of various types and degrees of kidney disease in diabetics has been esti- mated in population studies, using some clinical definition of "diabetic nephropathy' (11) but much more could be done. For example, the prognostic significance of so-called intermittent proteinuria and proteinuria of various degrees need clarification. It might be studied in well- defined geographically stable diabetic populations. Another closely related problem is the time course of the development of proteinuria and of albuminuria. Still another important clinical problem--one that must be solved on large groups of patients--is the question about "ethnic" differences in the development of diabetic angiopathy. It seems as if there are such differ- ences in the prevalence of various types of diabetes mellitus as defined by the degree of metabolic derangement (54), but there is still conflicting evidence about the prevalence of dia- betic retinopathy or diabetic nephropathy among Caucasians, American Indians, East Asians, etc., with comparable degrees of blood glucose abnormality. Editor's note: Investigations utilizing renal biopsies cannot be undertaken without full delibera- tion of the ethics of the proposed studies and the availability of alternate methods of assessment. There is a great need for noninvasive techniques for the evaluation of the earliest changes in para- meters of renal morphology and function. Hopefully, such methods will be developed and validated in parallel with the morphometric approaches. Alternatively, it may prove feasible to secure ade- quate information about renal vasculopathy by the less hazardous technique of muscle biopsy. Renal Disease 229 Such clinical studies will be particularly valuable if they specify and quantify certain degrees of renal insufficiency, of proteinuria, albuminuria, hypertension, edema, etc., in men and women, in certain age groups, and certain groups of age at the onset of diabetes. Here, as in all other long-range studies of diabetic vascular disease, results are much more likely to be informative and interesting if performed on patients with acute onset diabetes, preferably diagnosed before the age of 30 or 35 years. Diabetic nephropathy is a very common concurrent cause of death in long-term diabetic patients. However, it is well-known that the final stage of diabetic angiopathy is characterized by a combination of renal insuffiency, cardiac insuffiency, neuropathy (including diabetic encephalopathy) and evidence of diabetic macroangiopathy. It is often not possible to give a reasonable verdict as to the immediate cause of death in these patients. Moreover, the final picture of renal pathology itself is often a composite one. Ditscherlein (11) analyzed the histologic findings in 24 diabetics dying with the full-blown clinical picture of uremia. Glomerulosclerosis as the sole finding was observed in only eight cases. The other cases showed various combinations of diabetic glomerulosclerosis, chronic pyelonephritis, and nonspecific nephrosclerosis. Such studies are important and should be promoted when possible, but to be really significant they required a very high autopsy rate: ideally 100 percent. DIABETIC GLOMERULOPATHY Diabetic glomerulopathy appears at light microscopy as an increase in PAS-positive sub- stance. In mild cases this material is seen diffusely spread out within the glomerulus. In late stages it often forms the highly specific "Kimmelstiel-Wilson nodules." At electron microscopy glomerulopathy appears as a thickening of the peripheral basement membrane and an increase of the basement membrane-like material in the mesangial regions. (At light microscopy the basement membrane cannot be clearly distinguished). The basement membrane of the glomerular capillaries is analogous to that lining capillaries and larger blood vessels elsewhere in the body. A rather similar structure is found subjacent to other types of cells, e.g., epithelial cells, muscle cells, and interstitial cells. Thickening of the glomerular basement membrane is the local representative of the general process of the basement membrane thickening that occurs in all types of capillaries throughout the body. Only little information is available about the state of the basement membrane in the larger vessels in diabetic macroangiopathy. The same holds true for possible changes of non- vascular basement membranes. They will not be considered in the present chapter. The basement membrane of the renal glomerulus has been studied more intensively than that of any other vascular area. This is due especially to the fact that by its topography, it offers unique possibilities for exact morphological studies, and also that enough basement membrane glycoprotein is present in one kidney to allow isolation and biochemical studies. In comparison, the vascular areas of other organs present many problems. The retinal and iridal vessels cannot be biopsied. Quantitative studies of muscular capillaries are difficult for a number of reasons. However, it seems established by now--especially on the basis of the extended and excellent studies of Williamson and co-workers--that the onset and further develop- ment of vascular changes in these vessels are fundamentally similar to those occurring in the renal glomerulus (30,67). 230 Diabetes Mellitus Local factors undoubtedly play a role, e.g., vascular proliferations apparently only occurring in the retina and on the iris, but in looking for a general trend or common denominator of microvascular disease in diabetes mellitus, the glomerular vessels must be accepted at the moment as the most appropriate paradigm. THE GLOMERULAR BASEMENT MEMBRANE The basement membrane material which is involved in diabetic glomerulopathy is partly situated in the wall of the capillaries, here termed ''the peripheral basement membrane," and partly in the mesangial regions: "mesangial basement membrane-like material." Its shape, and its function as well differ at the two sites, but the ultrastructural appearance is practically identical. In contradistinction to earlier reports, it has now been shown that the development of abnormalities in diabetic patients occur at the same time at these two locations (72,73). The two separate morphological entities will therefore be considered together in the following. It has been known for many years that in the final stages of diabetic glomerulopathy, the tuft is nearly filled up with basement membrane material. With increasing duration of disease, there is increasing chance that severe involvement of the glomerular basement membrane will be found. By the use of quantitative electron microscopy, it has been possible to pinpoint exactly the earliest development of these abnormalities and, in particular, to date the relationship to the onset of the metabolic disease. This has been done on kidney biopsies from young patients with juvenile diabetes, i.e., patients in whom the time of onset is established with great certainty. It was shown that the glomerular basement membrane is normal at the acute onset of juvenile dia- betes (69). Studying the first few years after the diagnosis of diabetes, it was demonstrated that the amounts of basement membrane are increased already after 2 years, and to a greater ex- tent after 5 years (70,73). These initial abnormalities cannot be demonstrated with the light microscope, and at elec- tron microscopy only by means of extensive and rather cumbersome measurements. They represent the initial phases in a seemingly steadily progressive process, which finally leads to complete abolishment of glomerular function. The fact that the basement membrane is normal at the onset of diabetes has great practical and theoretical implications. It rules out, for all practical reasons, that the basement mem- brane thickening is a genetically determined abnormality--a view which has been popular for some years. This also implies that it may be possible, by controlling metabolic factors, to prevent or postpone the development of diabetic microangiopathy. Trials in which various clinical or therapeutic procedures are tested for the influence on the development of diabetic angiopathy are greatly needed, and as regards some specific questions, it is possible to carry them out. It has been documented that the thickness of the peripheral glomerular basement membrane is a useful parameter to evaluate the development when a precise quantitative technique is employed. The possibilities in this respect will be discussed later. THE GLOMERULAR CELLS What is the role and the fate of the glomerular cells in this development? This is an im- portant question, since the synthesis, the maintenance, and breakdown of the basement membrane are the result of cellular activity. Also the glomerular function as a filtering unit may be Renal Disease 231 modified at the cellular level. The established facts in this field are rather scanty. Most of the statements that have been put forward concern the number and differential distribution of the glomerular cells. Mesangial cell hyperplasia has been reported in light microscopic studies (27) and has been attributed a causative role in basement membrane overproduction. However, in a quantitative light-microscopic study of biopsies from patients with early diabetes (0-6 years' duration) cellular hyperplasia was not observed (74). The same conclusion was drawn from cell counts per- formed on electron microscopic photomontages of a smaller number of glomeruli from the same series of early diabetics. With this method, exact counts of individual cell types could be obtained, and it was found that the differential distribution of cells, as well as cell density, were normal (71). Cellular hyperplasia thus cannot account for the initial augmentation of base- ment membrane material. Obviously, very late in the development when glomeruli are about to undergo complete obso- lescence, the final outcome for the cells is total necrosis. What happens in between these two extremes of the development still needs to be studied with value quantitative techniques. The cellular function with respect to basement membrane synthesis and breakdown is not clarified. From silver-labeling experiments in animals, there is indication of basement membrane synthesis by epithelial and mesangial cells (34,35). It has not yet been possible to point out changes in cellular ultrastructure which could help to indicate the mechanism behind altered basement membrane metabolism in diabetes. However, cisternae of the endoplasmic reticulum of epithelial cells containing basement membrane-like material were observed with somewhat increased frequency in a series of young, short-term dia- betics compared to normals, although statistical significance was not obtained (71). There is some evidence that this organelle is a site of basement membrane synthesis (1). This ultra- structural finding is of interest in connection with the finding of elevated levels of enzymes involved in basement membrane synthesis as observed in alloxan-diabetic rats (58). Further studies at the ultrastructural level are very desirable. Such studies, dealing with the smallest details at a sub-ultrastructural level, may at first sight seem to have only small chance to lead to conclusions that may have practical consequences for the diabetic patients. However, some clues as to how to interfere with the abnormal mechanisms might be gained if some of the steps involved were elucidated. Possible ways to follow in an attempt to study cellular basement membrane synthesis at the ultrastructural level are: a) Quantitative determination on larger series of cases of cellular organelles involved in protein synthesis, i.e., the endoplasmatic reticulum with its cisternae, but also ribosomes, endoplasmic reticulum, Golgi apparatus. Such studies should preferentially run in parallel with biochemical determinations; b) Application of enzyme histochemistry--hope- fully demonstrating enzymes which are specific in the basement membrane turnover, using auto- radiographic techniques. MORPHOLOGY AND FILTER FUNCTION The functional consequences of basement membrane thickening are also unclarified. Since basal proteinuria and albuminuria are very late phenomena, basement membrane thicken- ing is present for many years before clinical signs of impaired filter function occur. Kidney 232 Diabetes Mellitus function as a whole depends on a complicated interplay between various segments of the nephron, including filtration, reabsorption, and secretion. Already immediately past the glomerular capillary wall, i.e., in the glomerular urinary space, the composition of the primary urine presumably is a result of a combined process of passive filtration over the basement membrane and reabsorption by glomerular epithelial cells. The basement membrane is believed to retain molecules larger than 100 A in diameter in normal individuals. Studies of the passage of dif- ferent enzymes indicated another barrier, the very fine membrane occluding the filtration slits between adjacent epithelial cells. This membrane was believed to stop somewhat smaller molecules with diameter down to about 65 A (36). However, in a recent publication on glomerular per- meability studied by means of dextrans of different molecular weights, the role of the filtra- tion slit membrane in the filtration process was drawn into doubt (8). It has been shown that a surface coat made up of mucopolysaccharides covers the epithelial cell membrane and probably fills out the filtration slits (25), thereby binding some of the substances passing and facili- tating their immediate reabsorption into the glomerular epithelial cell. In histochemical studies, a decrease in the glomerular sialic acid localized to this surface coat has been found in cases of proteinuria, both in human cases (4) and in animals with experimentally produced proteinuria (42). These findings might be related to a disappearance of the normal foot process structure. On the other hand, a thickening of the surface coat as measured in electron micro- graphs was reported in albuminuric rats (15). This coat of mucosubstances is not visualized in the electron micrographs when using conventional preparation of the tissue. It still remains to be clarified, with the application of special histochemical techniques, if abnormalities in these mucosubstances are present in diabetic patients and, if so, whether such abnormalities may account for alteration in the net result of the filtration process. In late cases of diabetic glomerulopathy, when gross proteinuria is present, the basement membrane must be anticipated to be leaky as are the thickened basement membranes of muscular, retinal, and iridal vessels. Still, its fine structure does not deviate markedly from the nor- mal, as far as we can determine with the presently available resolution. Since, however, it permits the passage of large quantities of protein molecules, the filter meshwork of these greatly thickened membranes must be altered in some way. Such sub-ultrastructural abnormalities could be evaluated with indirect techniques, studying the passage of tracer molecules of known sizes, which can be visualized on the electron micrographs. However, at present, such studies would have to be restricted to animal models, since techniques which could be used on kidney biopsies are not yet available. BIOCHEMISTRY OF THE GLOMERULAR BASEMENT MEMBRANE In the latter part of the 1960's, the foundation was laid for an understanding of the chemical nature of the basement nenbrane and its synthesis, by studies of material isolated from renal glomeruli. The results published in a long series of papers by Spiro and co-workers. Survey articles (56,58) may be briefly summarized as follows. The glomerular basement membrane is made up of a complex glycoprotein with a carbohydrate content of about 9 percent. The peptide portion contains large amounts of glycine and substan- tial amounts of hydroxyproline and hydroxylysine. Renal Disease 233 There are two distinct types of carbohydrate units, an unusual type of disaccharide unit made up of glucose and galactose and a heteropolysaccharide unit occurring also in other glycoproteins--made up of galactose, mannose, hexosamine, sialic acid, and fucose. The small unit is bound by glycoside linkage to hydroxylysine. Two enzymes involved in the synthesis of the glucose-galactose hydroxylysine, a glucosyltransferase, and a galactosyltransferase, were prepared, purified, and characterized. Studying glomeruli from patients who had been diabetic for 6 to 20 years, Beisswenger and Spiro (3) found that the content of hydroxylysine and hydroxylysine-linked disaccharide was increased. They suggested that the increase in disaccharide binding could reduce hypothetical aldol cross linkage of the chains, thereby resulting in increased permeability. In alloxan diabetic rats, they found an increase in the activity of the two trans- ferases which could be prevented by early insulin treatment. Many of these findings have been confirmed in other laboratories but a number of them have been disputed. There is controversy about the true nature of the normal basement membrane macromolecule, especially if real collagen is present in it. This is based, at least partly, on different opinions about how to separate the native subunits (29,55). More importantly in the present context, there is sharp controversy at the moment about possible differences between the base- ment membrane composition in normal and in diabetic subjects. Westberg and Michael (65) did not find any increase in hydroxylysine or hydroxylysine- linked disaccharide of glomerular basement membrane from diabetics, including patients who had had diabetes for more than 20 years. On the other hand, they observed a decrease in cystine content not reported by Beisswenger and Spiro (3). Westberg (64) suggests that this may signify a loosening of the interchange stability, perhaps resulting in increased permeability. Another finding of possible importance was a decrease in sialic acid in diabetic basement membranes. Kefalides (29) likewise found no difference in the amount of hydroxylysine linked dis- accharide of the glomerular basement membrane from normal subjects and patients with more than 10 years' duration of diabetes. In his study the cystine and sialic acid from diabetic basement membranes was low, as in the studies by Westberg and Michael (65). The high hopes that a biochemical explanation of basement membrane accumulation in dia- betes mellitus was already at hand have somewhat abated. Some of the confusing differences may be due to the fact that the techniques employed for separation, purification, and identification of small and large components of the basement membrane macromolecules have been different. Intensified efforts, if possible including close collaboration between individual labora- tories, are highly desirable in order to solve some of the problems that have arisen in this potentially very significant field of diabetes research. The limitation of biochemical basement membrane studies is that it does not provide infor- mation specifically about the metabolism of the glomerular basement membrane producing cells. An approximation to this aim is attained in the interesting study of Cohen (9), demonstrating inter alia an increased lysine incorporation in glomerular basement membrane and in non- dialyzable protein of subcellular fractions. However, combined morphological and biochemical studies are required to produce information about the metabolic processes in the individual cell types, especially in the epithelial cell. 234 Diabetes Mellitus Finally, although perhaps not to be expected at the moment, the possibility of biochemical studies on human biopsy material should be mentioned. However, this will require a degree of micromanipulation and microanalysis not available at the moment. KIDNEY FUNCTION IN EARLY AND LONG-TERM DIABETES, AND PROTEINURIA It has been shown many years ago and confirmed recently that renal plasma flow, glomerular filtration rate, and various tubular functions are reduced to about the same degree in long-term diabetes with pronounced morphological abnormalities of the renal tissues. See review article (39). The fact, recently studied in much detail (12,45) that various functional parameters show high values in recent diabetes is surprising and cannot be fully explained today. The '"hyper- function" is not related directly to the momentary blood glucose level but is reversible in the sense that it can be normalized after some time of strict regulation of the metabolic state (45). Macromolecular clearance studies of the effectiveness of the filtering structures failed to demonstrate any increase in functional pore size (43). Increase in filtration pressure may play a role since high filtration fraction has consistently been found in early juvenile diabetes. Another possible explanation would be an increase in filtering area. It has been shown recently that the roentgenographic kidney size is increased to the same extent as is the GFR in early diabetes (46). After 3 months of strict insulin treatment, the kidney size and GFR had decreased to near normal values (47). Quantitative light-microscopical studies of individual glomeruli made it possible to demonstrate the the glomerular size is increased (74). It is of special interest that the volume of the capillary lumen is enlarged. This finding may well be the morphological counterpart of the increased GFR, and it may be due to increased filtration surface. For technical reasons the filtration surface could not be evaluated in the kidney biopsy material available for this study. The mechanism of this enlargement may involve the high level of plasma growth hormone characterizing diabetes mellitus when not exceedingly well controlled (18,19). In animal models it may be possible to correlate such morphological parameters with the degree of metabolic disturbance in the untreated state as well as after some time of insulin treatment. The clinical importance, if any, of all this is not clear. Looking at the development during the whole life span of the diabetic patients, the situation can be described metaphor- ically as a conflict between a metabolically determined tendency towards high function and an angiologic tendency towards low function, the angiologic one finally winning the upper hand. It has been suggested that the hyperfunction of the early period may play a role in modulating the transition to low function in the late phase (12). Proteinuria is often seen in long-term diabetes and heavy proteinuria occurs in the final stage in some cases. Editor's note: Methodological difficulties have restricted studies of the normal biology and bio- chemistry of the specific cell types affected in diabetic nephropathy. Recent reports that human umbilical vein endothelial cells can be maintained and studied in tissue culture (Jaffe, E. A., R. L. Nachman, C. G. Becker, and C. R. Minick: Culture of Human Endothelial Cell Derived from Umbili- cal Veins. J. Clin. Invest. 52:2745, 1973) give hope that similar studies can be performed with renal mesangial cells. If successful they may alter the opportunities for progress in this field. Renal Disease 235 This phenomenon is usually regarded as an expression of leakage of the filtering structures (it might also be due, partly or totally to decreased tubular reabsorption). There is no contradiction between the finding of a generally thickened basement membrane and the idea of an increased passage of protein through it. A loosening of the cross-linking of the long-chain molecule structure of the thickened basement membrane is quite possible, although it has not been demonstrated with current electronmicroscopic techniques. Functionally it has been shown that there is an increased permeability of the thickened basement membrane of muscular capil- laries (60) and also of retinal (23) and iridal vessens (2,24). One interesting new finding that needs to be further investigated is the seeming difference between proteinuria, as determined with the Lowry method and albuminuria, determined with radio- immunological techniques. Protein excretion is reported to increase steadily during the years (52). The excretion of albumin is normal and remains normal in reasonably well controlled dia- betics till after many years of diabetes when suddenly, so it appears, it begins to rise (44). However, using physical exercise as a provocative test (work load kpm/min for 20 minutes) a clearcut abnormality in albumin excretion emerges in diabetics with a duration of diabetes of only a few years--i.e., at the point of time when morphological changes are known to be present but basal albumin excretion is still normal. This work load results in a pronounced rise of albumin excretion in the diabetics, while normal subjects are not affected (48,49). PYELONEPHRITIS AND BACTERIAL INFECTION OF THE RENAL TISSUE It was thought for many years that clinical symptoms of infectious kidney disease were common in diabetic patients. Based on histologic findings and autopsy, it was also accepted that chronic pyelonephritis was more common in diabetics than in nondiabetics (26). There is no doubt, of course, that what has been described for half a century as ''chronic pyelonephritis" is a histologic reality, but recent critical studies have failed to establish that this picture is caused by chronic bacterial infection (32). First, it was pointed out that the histological picture of "chronic pyelonephritis" could just as well be caused by glomerulosclerosis and/or by impaired blood supply due to the more or less well-defined changes of large and medium size renal blood vessels supposedly more common in diabetics than in nondiabetics. Second, attempts to cultivate bacteria from biopsy-specimens obtained in cases of so-called chronic pyelonephri- tis have usually been unsuccessful (6,17). These problems are not limited to chronic pyelonephritis in diabetes mellitus. They are more provocative, however, in diabetes, because there is no doubt that urinary tract infection, as defined below, is more common in diabetics than in nondiabetics. It is not easy to indicate ways out of the present crisis of ''chronic pyelonephritis' as a chronic renal tissue infection. The first point mentioned above could be studied again in greater detail and in larger series of diabetics and nondiabetic cases meeting the former admittedly rather vague histological criteria of 'chronic pyelonephritis.’ The elucidation of the second point is hampered by the fact that this abnormality, whatever its nature, is certainly mostly found as small irregularly distributed focal processes. Continued efforts of finding bacteria in biopsy specimens is hardly an attractive suggestion. Recent experimental and clinical studies have indicated, however, that raised titres to urinary tract bacterial antigens are found in a certain number of upper urinary tract infections (22,53). It may be worthwhile to apply these techniques to the problem of renal tissue infection in diabetic patients. 236 Diabetes Mellitus Considering acute pyelonephritis the histologic diagnosis is much more clear-cut. It has been found in large autopsy series that signs of acute infections in the kidney tissue do occur more frequently in diabetics than in controls (11). In this connection the rather high fre- quency of papillary necrosis in diabetics should also be mentioned (37). Although the etiology of this syndrome is not entirely clarified, it seems clear that both infection and impaired blood supply to the renal medulla are important factors in some cases. The clinical study of urinary tract infection was greatly promoted by the introduction of quantitative techniques for the evaluation of bacteriuria distinguishing between contamination and significant bacteriuria (i.e., > 10° bacteria per ml urine). Applying this method to dia- betic and nondiabetic populations, results were at first presented which contradicted the state- ment of increased frequency of urinary tract infections in diabetics (51). However, in later studies a higher frequency of significant bacteriuria in diabetic patients was clearly demon- strated, at least in females (61,68). Significant bacteriuria was shown, moreover, to be positively correlated with the duration of diabetes and with the occurrence of retinopathy--thus by implication also with the presence of diabetic glomerulopathy (62). If this coexistence is an expression of decreased resistance to infections in a kidney which is the seat of generalized vascular involvement, or if recidivat- ing infection may accelerate the development of diabetic angiopathy in the kidney is not clear, although the first-mentioned possibility is more likely to be correct. At any rate, it is clear that acute, recurrent infections within the kidney tissue may be contributory to the destruction of renal parenchyma, and may thereby be a significant co- factor in the development of renal insufficiency. Therefore, there is a great need for further studies in this field. HYPERTENSION IN DIABETES It is often thought that arterial hypertension in diabetes mellitus, or in some diabetics, is in some way connected with diabetic kidney disease. It may be so but there is no strong evidence for it. Moreover, there is no unanimity as to the simple question about the prevalence of arterial hypertension in diabetes. Partly, at least, this confusion is due to the fact that many investigators have neglected the two aspects known to be significant in most areas of dia- betological angiology, viz. metabolic state and duration of diabetes. Experimental studies have shown that blood pressure is slightly but demonstrably elevated in resting and exercising diabetic patients in poor state of control as compared to well- controlled diabetics or normals (7,16). The 5-10 mm higher average blood pressure found in a group of children after less than 15 years of diabetes, when compared with nondiabetic children, may have been due to incomplete normalization of the metabolic state (50). A much quoted study of a large population showed that there was no statistically significant difference between the blood pressure in diabetics and in nondiabetics when age and sex were taken into account (14). Unfortunately, the relationship to the duration of diabetes was not analyzed. It has been shown, however, that the blood pressure of juvenile diabetics increases Editor's note: The difficulties in differentiating between bacterial involvement of the upper and lower urinary tract have long plagued the clinician. The recent application of immunofluroescent techniques to localize the source of urinary bacteria on the basis of their coating with antibody has facilitated diagnostic discrimination (Jones, S. R., J. W. Smith, and J. P. Sanford: New Engl. J. Med. 190:591, 1974; Sanford, J. P.: Ann. Rev, Med. 26:485, 1975). Renal Disease 237 with the duration of diabetes (66). In an unselected and representative series of long-term diabetics of all ages (duration of diabetes 15-25 years), arterial hypertension (systolic blood pressure > 150, diastolic > 100) was found in 18 percent of patients aged less than 40. In men and women more than 50 years old, it occurred in 84 percent (38). These values are obviously higher than those occurring in the general population and may well be connected with renal disease. In the study quoted there was a statistically significant correlation between hyper- tension and abnormal urine sediment. The clinical importance of hypertension per se in diabetes mellitus is doubtful. Recent studies, contradicting earlier opinions, have indicated that the frequency of acute cerebro- vascular accidents is higher in diabetics than in nondiabetics (13). It is not clear, however, how much of this pathology is due to arterial hypertension and how much to large or small vessel abnormalities independent of blood pressure. In the initial phase of clinical diabetic nephropathy, i.e., when the patients begin to develop proteinuria, the blood pressure is often somewhat elevated. In such patients glomerular filtration rate decreases gradually in the course of years and the rate of this fall is cor- related to the diastolic blood pressure. Antihypertensive treatment may therefore be of value at this stage. It has been shown that correction of the mild hypertension in such patients is followed by a decrease of the albumin excretion (47). Further studies will clarify whether antihypertensive treatment can also postpone the development of renal insufficiency in patients with diabetic nephropathy. Severe arterial hypertension is uncommon in diabetes. Even in the series of long-term dia- betics mentioned above, a diastolic pressure above 120 was seen in only 6 percent of the cases. This statement requires, however, one modification: it is general clinical experience that blood pressure is often considerably elevated in the last weeks or months of life of patients succumb- ing to diabetic nephropathy and other manifestations of diabetic angiopathy. This fact is well documented in the study of Watkins et al (63). THERAPY AND PROPHYLAXIS Diabetic glomerulopathy is an element of diabetic angiopathy in general. This has been clear for many years from the resemblance between light microscopic changes in the glomerulus and in other organs of the body, as well as from the well established statistical connections between the prevalence and degree of vascular changes in the kidney, the eye, the heart, etc. (38,59). The problems of therapy and prophylaxis of diabetic glomerulopathy are therefore partly identical with the problems of therapy and prophylaxis of diabetic angiopathy in general. These general problems will not be discussed in detail here, but the usefulness of the results ob- tained from kidney studies deserve to be mentioned. The recent developments in quantitative electronmicroscopic analysis of diabetic glomerul- opathy described above have opened up new possibilities for estimating the effect of old as well as new therapeutic/prophylactic measures in diabetic renal disease as an expression of diabetic angiopathy in general. Earlier studies were hampered by the many years it takes for diabetic vascular disease to appear clinically and to develop to severe forms, and also by the crude means of estimating appearance and development of vascular changes. 238 Diabetes Mellitus The fact that thickening of glomerular basement membrane can be detected after only 2-3 years makes it reasonable to envisage controlled clinical trials to determine the effect of any therapeutic or prophylactic measure, all the way from ''good control' to new types of insulin, or, hopefully, pharmacological suppression of the overproduction of growth hormone in diabetic patients. Trials could be constructed in which groups of young diabetic patients were allocated to one of two alternative treatments such as ''good control" versus 'very good control,' ordinary insulin versus monocomponent insulin, etc. Growth hormone having been proposed as a causal factor in the development of diabetic vascular disease (40,41), a trial of the effect of growth hormone suppression is indicated. The newly discovered hypothalamic growth hormone suppressor, somatostatin, provides the means of pharmacological and reversible suppression of growth hormone (5,20). At the present time, how- ever, such a test is not feasible because somatostatin is an extremely short-acting compound and also because of the many as yet not fully understood actions of this preparation on various pituitary and extra-pituitary hormones (21). The manufacture of a long-acting somatostatin analog with action confined more exclusively to growth hormones will be necessary before con- trolled clinical trials can be initiated to test its effectiveness in inhibiting the development of diabetic renal disease and other vascular abnormalities in diabetic patients. It should be possible and ethically acceptable to ask a reasonable number of informed young patients and their families to participate in controlled clinical trials covering a period of only 2-3 years. The group of patients turning out in the end to have had the relatively un- favorable treatment could then be switched over to the favourable regimen at a point of time when vascular changes were minimal, i.e., only just recognizable with the electronmicroscopic technique described above. There is no specific prophylaxis or treatment of diabetic nephropathy as such. The treat- ment of chronic renal infection or pyelonephritis in diabetes does not differ from that in non- diabetics. It is equally unsatisfactory. The same was true for renal insufficiency until recently. Hemodialysis and renal transplantation are now being performed on diabetic patients in certain centers (33). This raises many problems. First of all, there is a lack of information to elucidate the number of patients per year in a given area or country who could be considered as candidates for these procedures. In places where cases of uremia are registered, it should be possible to find the uremic diabetics in order to ascertain the incidence and degree of ocular disease, heart disease, and nervous system disease. In this way an estimate may be ob- tained of the benefit that would accrue from a systematic introduction of modern therapy in long-term diabetics with renal insufficiency. It should be noted, however, that long-term re- sults are not yet available elucidating the fate of a successfully transplanted kidney in the body of a patient with severe diabetic angiopathy, especially if and when the diabetic glomerulopathy will appear in it. Renal Disease 239 REFERENCES 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. Andres, GA, C Morgan, KC Hsu, RA Rifkind, and BC Seegal 1962. Electron microscopic studies of experimental nephritis with ferritin-conjugated antibody. The basement membranes and cisternae of visceral epithelial cells in nephritic rat glomeruli. J Exp Med 115:929-935. Baggesen, LH 1969. Fluorescence angiopathy of the iris in diabetics and nondiabetics. Acta Ophthal Kbh 47:449-460. Beisswenger, PG, and RG Spiro 1972. Studies on the human glomerular basement membrane. Composition, nature of the carbohydrate units and chemical changes in diabetes mellitus. Diabetes 22:180-193. Blau, EB, and JE Haas 1973. Glomerular sialic acid and proteinuria in human renal disease. Lab Invest 28:477-481. Brazeau, P, W Vale, R Burgus, N Ling, M Butcher, J Rivier, and R Guillemin 1973. Hypotha- lamic polypeptide that inhibits the secretion of immunoreactive pituitary growth hormone. Science 179:77-79. Brun, C, and F Raaschou 1961. Percutaneous renal biopsy in pyelonephritis. In GEW Wolstenholme, and MP Cameron, eds. Renal biopsy: clinical and pathological significance. Boston, Little, Brown, p 245. Carlstrom, S, and T Karlefors 1967. Hemodynamic Studies during Exercise in Newly Diagnosed Juvenile Diabetics. Acta Med Scand 181:759-767. Caulfield, JP, and MG Farquhar 1974. The permeability of glomerular capillaries to graded dextrans. Identification of the basement membrane as the primary filtration barrier. J Cell Biol 63:883-903. Cohen, MP 1973. Renal glomerular basement membrane synthesis in experimental diabetes. Abstracts, 8th Congr Internat Diab Fed, Excerpta Med, Amsterdam, p 174. Danowski, TS, ER Fisher, RC Khurana, S Nolan, and T Stephan 1972. Muscle capillary base- ment membrane in juvenile diabetes mellitus. Metabolism 21:1125-1132. Ditscherlein, G 1969. Nierenveranderungen bei Diabetikern. Jena, VEB Gustav Fischer Verlag. Ditzel, J, and K Junker 1972. Abnormal glomerular filtration rate, renal plasma flow, and renal protein excretion in recent and short-term diabetics. Brit Med J 2:13-19. Entmacher, PS, HF Root, and HH Marks 1964. Longevity of diabetics in recent years. Diabetes 13:373-377. Freedman, P, R Moulton, and AG Spencer 1958. Hypertension and diabetes mellitus. Quart J Med New Ser 27:293-305. Groniowski, J, W Biczysko, and M Walski 1974. Electron microscopic studies on the surface coat of renal podocytes in albuminuric rats. Lab Invest 30:58-63. Gundersen, HJG 1974. Peripheral blood flow and metabolic control in juvenile diabetes. Diabetologia 10:225-31. Halverstadt, DB, GW Leadbetter, and RA Field 1966. Pyelonephritis in the diabetic. Correlation of open renal biopsies and bacteriological studies. JAMA 195:827-829. Hansen, AgP 1972. Serum grown hormone patterns in juvenile diabetes. Dan Med Bull 19, suppl 1. Hansen, AaP, and CE Morgensen 1972. Growth hormone secretion and kidney function during normalization of the metabolic state in newly diagnosed diabetes. Horm Metab Res 4:11-15. 240 Diabetes Mellitus 20. 21. 22 23. 24. 25. 26. 27. 28. 29. 30. 31. 22. 33. 34. 35. 36. 37. 38. 39. 40. Hansen, AaP, H @rskov, K Seyer-Hansen, and K Lundbae k 1973. Some actions of growth hormone release inhibiting factor. Brit Med J 3:523-524. Hansen, AaP, K Lundba k, CH Mortimer, GM Besser, R Hall, and AV Schally 1975. Growth hormone release inhibiting hormone: its action in normals, acromegalics and diabetics. International Society of Neuroendocrinology, Serono Symposia. Hypothalamic Hormones: Chemistry, Physiology, Pharmacology and Clinical use, Milano. (In press.) Heptinstall, RH, and PW Ramsden 1972. Antibody production in urinary tract infections in the rat. Invest Urol 9:426-430. Hodge, RV, and CTD Dollery 1964. Retinal soft exudates. Quart J Med (N.S.) 33:117-130. Jensen, VA, and K Lundbae k 1968. Fluorescence angiography of the iris in recent and long-term diabetes. Diabetologia 4:161-163. Jones, DB 1969. Mucosubstances of the glomerulus. Lab Invest 21:119-125. Joslin, EP, HF Root, P White, A Marble, and CC Bailey 1946. The treatment of Diabetes Mellitus. 8th Ed London. Kawano, K, M Arakawa, J McCoy, J Porch, and P Kimmelstiel 1969. Quantitative study of glomeruli. Focal glomerulonephritis and diabetic glomerulosclerosis. Lab Invest 21: 269-275. Kefalides, N 1973. Discussion, in (Eds) Camerini-Ddvalos & Cole, H.S.: Vascular and neurological changes in early diabetes. Aca. Press, New York and London, p. 189. Kefalides, NA 1974. Biochemical properties of human glomerular basement membrane in normal and diabetic kidneys. J Clin Invest 53:403-407. Kilo, C, N Vogler, and JR Williamson 1972. Muscle capillary basement membrane changes related to aging and diabetes mellitus. Diabetes 21:881-905. Kimmelstiel, P, and C Wilson 1936. Intercapillary lesions in the glomeruli of the kidney. Am J Path 12:83. Kimmelstiel P, OJ Kim, JA Beres, and K Wellmann 1961. Chronic pyelonephritis. Am J Med 30:589-607. Kjellstrand, CM, RL Simmons, FC Goetz, TS Buselmeier, JR Shideman, B Hartitzsch, and JS Najarian 1973. Renal transplantation in patients with insulin-dependent diabetes. Lancet 2:4-8. Kortge, P, J Schurholz, and A Scholl 1969. Uber den Beitrag der Mesangiumzellen und der Glomerulumdeckzellen an der Bildung der Basalmembran der Glomerulumkapillaren. Verh Deutsch Gesell d Pathol 53:384-390. Kurtz, SM, and JD Feldman 1962. Experimental studies on the formation of the glomerular basement membrane. J Untrastruct Res 6:19-27. ° Latta, H 1970. The glomerular capillary wall. J Ultrastruct Res 32:526-544. Lauler, DP, GE Schreiner, and A David 1960. Renal medullary necrosis. Am J Med 29:132- 156. Lundba k, K 1953. Long-term diabetes--the clinical picture in diabetes mellitus of 15-25 years duration, with a follow-up of a regional series of cases. Monograph. Munksgaard Editor, Copenhagen. Lundbae k, K 1973. Recent advances in the study of diabetic renal disease. Pathobiology Annual 3:377-404. Lundba k, K, NJ Christensen, VA Jensen, K Johansen, TS Olsen, AP Hansen, H @rskov, and R @sterby 1970. Diabetes, diabetic angiopathy and growth hormone. Lancet II:131. 41. 42, 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60, 61. Renal Disease 241 Lundbae k, K, NJ Christensen, VA Jensen, K Johansen, TS Olsen, AP Hansen, H @rskov, and R @sterby 1971. The pathogenesis of diabetic angiopathy and growth hormone. Dan Med Bull ‘18:1. Michael, AF, E Blau, and RL Vernier 1970. Glomerular polyanion. Alteration in amino- nucleoside nephrosis. Lab Invest 23:649-657. Mogensen, CE 197la. Kidney function and glomerular permeability to macromolecules in early juvenile diabetes. Scand J Clin Lab Invest 28:79-90. Mogensen, CE 1971b. Urinary albumin excretion in early and long-term juvenile diabetes. Scand J Clin Lab Invest 28:183-193. Mogensen, CE 1972. Kidney function and glomerular permeability to macromolecules in juvenile diabetes. Dan Med Bull, Suppl. Mogensen, CE, and MJF Andersen 1973. Increased kidney size and glomerular filtration rate in early juvenile diabetes. Diabetes 22:706-712. Mogensen, CE, and MJF Andersen 1975. Increased kidney size and glomerular filtration rate in untreated juvenile diabetes. Normalization by insulin treatment. Diabetologia, in press. Mogensen, CE, and E Vittinghus 1975a. Urinary albumin excretion during exercise in juve- nile diabetes. A provocation test for early abnormalities. Scand J Lab Clin Invest (in press). Mogensen, CE, E Vittinghus, and K S@#lling 1975b. Increased urinary albumin, light chain and B-2-microglobulin excretion in normal man. (To be published.) Moss, AJ 1962. Blood pressure in children with diabetes mellitus. Pediatrics 30:932-936. O'Sullivan, DJ, MG Fitzgerald, MJ Meynell, and JM Malins 1961. Urinary tract infection. A comparative study in the diabetic and general populations. Brit Med J 1:786-788. Panzram, G, G Unger, and H Wollner 1967. Untersuchungen tuber physiologische Proteinurie beim Diabetes mellitus. Dtsch Med Wschr 92:1013-1016. Percival, A, W Brumfitt, and J deLouvois 1964. Serum-antibody levels as an indication of clinically inapparent pyelonephritis. Lancet II1:1027-1033. Rushforth, NB, PH Bennett, AG Steinberg, TA Burch, and M Miller 1971. Diabetes in the Pima Indians. Diabetes 20:756-765. Spiro, RG 1970. Chemistry and metabolism of the basement membrane. In M Ellenberg, and H Rifkin, eds. Diabetes mellitus: Theory and Practice. New York, McGraw-Hill, p 210. Spiro, RG 1972. Basement membranes and collagens. In (ed) A Gottschalk: Glycoproteins, vol 5B, Elsevir, Amsterdam, pp 965-999. Spiro, RG 1973. Discussion. In (eds) Camerini-Ddvalos, and HS Cole: Vascular and neurological changes in early diabetes. Acad Press, New York and London, p 188. Spiro, RG, and MJ Spiro 1971. Effect of diabetes on the biosynthesis of the renal glomerular basement membrane. Diabetes 20:641-648. Thomsen, AaChr 1965. The kidney in diabetes mellitus. Munksgaard, Copenhagen. Trap-Jensen, J, and NA Lassen 1968. Increased capillary diffusion capacity for small ions in skeletal muscle in long-term diabetics. Scand J Clin Lab Invest 21:116-122. Vejlsgaard, R 1966a. Studies on urinary infection in diabetics. I. Bacteriuria in patients with diabetes mellitus and in control subjects. Acta Med Scand 179:173-182. 242 Diabetes Mellitus 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. Vejlsgaard, R 1966b. Studies on urinary infection in diabetics. II. Significant bacteriuria in relation to long-term diabetic manifestations. Acta Med Scand 179:183-188. Watkins, PJ, JD Blainey, DB Brewer, MG Fitzgerald, JM Malins, DJ O'Sullivan, and JA Pinto 1972. The natural history of diabetic renal disease. Quart J Med, New Series, 41:437-456. Westberg, NG 1973. Human glomerular basement membrane. Composition in normal and de- creased human kidneys, with special reference to diabetic nephropathy. Thesis, University of Goteborg, Sweden. Westberg, NG, and AF Michael 1973. Human glomerular basement membrane: Chemical composi- tion in diabetes mellitus. Acta Med Scand 194:34-47. White, P 1956. Natural course and prognosis of juvenile diabetes. Diabetes 5:445-450. Williamson, JR, NJ Vogler, and C Kilo 1973. Basement membrane thickening in muscle capillaries. Observations on diabetics and nondiabetics with both parents diabetic. Proceedings of the IVth International Congress on Endocrinology. Internat. Congress Series No 273, pp 1122-1128. Excerpta Medica, Amsterdam. @sterby Hansen, R 1964. Bacteriuria in diabetic and nondiabetic out-patients. Acta Med Scand 176:721-730. ’ @sterby Hansen, R 1965. A quantitative estimate of the peripheral glomerular basement membrane in recent juvenile diabetes. Diabetologia 1:97-100. @sterby, R 1972a. Morphometric studies of the peripheral glomerular basement membrane in early juvenile diabetes. I. Development of initial basement membrane thickening. Diabetologia 8:84-92. @sterby, R 1972b. The number of glomerular cells and substructures in early juvenile dia- betes. A quantitative electron microscopical study. Acta Pathol Microbiol Scand A 80: 785-800. i @sterby, R 1973. A quantitative electron microscopic study of mesangial regions in glomeruli from patients with short-term juvenile diabetes mellitus. Lab Invest 29:99-110. @sterby, R 1975. Early phases in the development of diabetic glomerulopathy. A quantita- tive electron microscopic study. Acta Med Scand, suppl 574. @sterby, R, and HJG Gundersen 1975. Glomerular size and structure in diabetes mellitus. I. Early abnormalities. Diabetologia 11, in press. 18 DIABETIC PERIPHERAL NEUROPATHY J. A. Moorhouse INTRODUCTION Peripheral neuropathy is the earliest and most universal recognizable complication of diabetes mellitus. Morphologic (68), electrophysiologic (44, 46) and sensory-perceptive (12) abnormalities are present within days to weeks of its onset. Even this early neuropathy is not limited to sen- sation in the lower extremities. The special senses, finger-tip perception, and motor innervation are affected (12,52). These and other observations appear to resolve the question of whether neuropathy is a compli- cation or a separate genetic concomitant (31, 33) of the diabetic state. Electrophysiologic and sensory-perception abnormalities are not present in the nondiabetic immediate relatives of dia- betic persons (12, Moorhouse, unpublished observations). Abnormalities in individuals with mild carbohydrate intolerance are absent or slight (15, 35). The morphologic (49, 76) and electro- physiologic (28, 40) manifestations of diabetic neuropathy regularly occur in nongenetic experi- mental diabetes. The electrophysiologic manifestations of diabetic neuropathy are readily corrected by early insulin treatment (41,45,46). These findings make it implausible that diabetic neuropathy has a primary genetic basis rather than being the consequence of disordered metabolism. I An overall question which does require resolution is whether primary damage to the nervous system in diabetes is multicentric, or whether a single target site can be defined. There is nothing in the existing literature which is inconsistent with the view that peripheral satellite cells of Schwann are the primary targets for damage in diabetes, and that the subsequent degen- erative sequence is secondary to this primary lesion. PATHOLOGY AND PATHOGENESIS Diabetic peripheral neuropathy is characterized structurally by segmental demyelination (14, 82). Occlusive lesions of the vasa nervorum usually are minimal even in advanced neuropathy. Axons are normal to light microscopy until the disease is well advanced. Early nonspecific axonal abnormalities are visible to electron microscopy (3.7.8). The possibility that they are primary needs to be excluded. Segmental demyelination is due to damage to the supporting cells of Schwann which surround a all peripheral nerve fibers. The myelin sheath is derived embryologically from the plasma mem- brane of the Schwann cell (5) and continues to be dependent upon that cell for preservation and support. In diabetes early degenerative changes within the Schwann cell (3,7) indicate it to be a primary target for damage. Myelin Sheath Splitting and separation of the myelin lamellae at paranodal regions is the earliest recog- nizable change (3,82). The degenerative process may advance to complete segmental loss, or may at any stage be followed by healing. Insulin replacement promotes healing (44, 45), but even without insulin the changes have a natural cyclicity which is of interest both in relation to 243 244 Diabetes Mellitus diabetes and to Schwann-cell physiology. Myelin regeneration tends to be disorderly and irregular (14). Schwann Cell Electron microscopy of the Schwann cell reveals lamellar lipoprotein inclusions and other amorphous inclusions which reflect a disturbance of its normal synthesizing function (3,7). Other changes are abnormal glycogen granules and thickening and reduplication of the cell basement membrane. These changes are present in nerve biopsy specimens from all diabetic patients from early in the disease. In a given biopsy specimen, however, their occurrence is markedly nonuniform. Only a minority of Schwann cells show abnormalities, and these vary from slight to severe. The factor(s) causing these intercell differences is unknown. Light-microscopic changes occur much later. There develops a nonspecific hypertrophy of Schwann cells, which eventually occur in the form of clusters (82). This picture is seen in any disorder which results in repeated degeneration and regeneration of the myelin sheath. Axons In early diabetes axons are normal in number and appearance to light microscopy (14), although with electron-microscopy they may show increased cytoplasmic dense bodies and disorderly arrange- ments of neurofilaments (7, 8). These changes are nonspecific and probably are secondary to damage to the satellite cells, although the possibility that they are primary needs to be excluded. The axon, extending from its cell body over great distances, is dependent upon its satellite cells for local metabolic energy and control (73, 77). Nerve Roots Studies in early diabetes are not available. In late cases the dorsal roots are often severely demyelinated (24). It is of interest that in contrast the anterior roots may be normal to light microscopy. This observation, which could be sought in early experimental diabetes, might provide investigators with considerable leverage. At this paraspinal site, motor and sensory fibers are conveniently separated. What ultramicroscopic or biochemical differences characterize those fibers susceptible and those less susceptible to metabolic damage in diabetes? Dorsal-Root Sensory Ganglia Lesions in the sensory ganglia are minor and variable until late in diabetes (9,24), and even then are nonspecific. There occurs proliferation of ganglion-capsule cells and of fine nonmy- elinated nerve fibers, with compression-degeneration of the ganglion cell-bodies. These changes occur in many neuritides and even after posterior rhizotomy. Therefore they are not in conflict with a unitary, Schwann-cell hypothesis for the pathogenesis of diabetic neuropathy. By that hypothesis, they are tertiary to Schwann-cell damage, secondary to axonal degeneration. Spinal Cord It is remarkable that oligodendrocytes, the axon-satellite cells of the central nervous sys- tem which fulfill the same function as the Schwann cells in the peripheral nervous system, are immune to discernible damage at least until late in the course of diabetes (9,27,67). Normal central myelin is similar in ultramicroscopic apperance to peripheral myelin, although it differs in lipid and in aminoacid composition (20). What crucial difference in the properties of the Schwann cell and of the oligodendrocyte makes the one so vulnerable to damage in diabetes and the other so impervious? } Fr To light microscopy the spinal cord frequently is normal in diabetes. In advanced diabetes, as well as infarcts, there may occur irregular swellings of axons and of myelin sheaths, and Diabetic Peripheral Neuropathy 245 varying degrees of axon loss, most commonly in the posterior columns of the lumbosacral region. These findings are nonspecific, and their lateness and distribution suggests that they are secon- dary to peripheral nerve damage. Similar changes may follow posterior rhizotomy. Therefore they are not in conflict with a unitary hypothesis for diabetic neuropathy. Cranial Nerves There are no observations in early diabetes. In advanced diabetic neuropathy the peripheral portions of cranial nerves may show severe demyelination (70). Again it is arresting that in such cases the proximal portions, where oligodendrocytes replace Schwann cells as axon satellites, showed only some swelling and inequality of the myelin sheaths. This juxtaposition of Schwann cells and oligodendrocytes along the same fibers might enable an ultramicroscopic and biochemical analysis of the differences which affect their susceptibility to damage in diabetes. Brain Frequently the brain is normal to light microscopy in diabetes (9,24,70). In late cases variable degenerative changes may occur both in grey and white matter. It is not clear whether these changes are to any extent characteristic of diabetes, or whether they are only the results of hypertension and of atherosclerosis occurring at an earlier age than in nondiabetic subjects. Even if they are to some extent characteristic, they do not appear to be inconsistent with a unitary hypothesis, with late transynaptic extension of damage to neuronal systems having close functional links with those primarily affected. Autonomic System Nonmyelinated axons in peripheral nerves are closely invested by Schwann cells. The changes in diabetes seen by electron microscopy are similar to those in myelinated fibers (3, 8). Indeed in early diabetes the nonmyelinated fibers are quantitatively the more affected. As the disorder progresses, nonmyelinated axons become depleted in peripheral nerves (24,82). Studies of sympathetic ganglia and trunks in diabetes are scarce and limited to light micro- scopy. Ill-defined degenerative changes are described within ganglion cells and in the rami com- municantes (1,9,48). Further morphologic and functional studies are needed both in clinical and in experimental diabetes. Changes in the vagus nerve are said not to be present to light microscopy of clinical material (48), but this is belied by the occurrence of esophageal hyperresponsiveness to stretch stimuli even in early diabetes (50), and by impaired gastric secretory responses to hypoglycemia in more advanced neuropathy (25). There is no information about the appearance of gastric intramural gan- glia and plexuses (86). there are said to be no definite changes in intestinal intramural ganglia and plexuses (24, 68). Degenerative changes are said frequently to be present in bladder intra- mural fibers and in hypogastric nerve fibers (34). The morphology and function of the parasympa- thetic system in diabetes requires much further study. Sensory Nerve Endings Quantitative studies of peripheral sensory nerve endings suggest that their number is dimin- ished in various forms of neuropathy, including that of diabetes (22, 27). Confirmatory duration- specific studies are needed, as well as electron microscopic observations on sensory-ending ultrastructure. Motor End Plates Morphologic abnormalities of the terminal neuromuscular apparatus are uniformly present in 246 Diabetes Mellitus muscle biopsies (2,16,69). They occur within weeks of the onset both of clinical and of experi- mental (49) diabetes. They persist throughout the course of the disease in degeneration-regeneration cycles with degeneration eventually becoming ascendent in patients with severe neuropathy. Their etiology is unknown, but it is noteworthy that the terminal branches of motor nerves are covered with a thin sheath of Schwann cells right up to their end swelling (3). Electron micro- scopy of motor nerve terminals in diabetes has not been performed. The histopathologic changes begin with the appearance of motor end plates which are irregular in size and shape. Whether or not degenerative changes in muscle fibers occur at this state is unclear (16, 68). This end-plate dystrophy is followed by the appearance of degenerative frag- mentations and swellings along terminal nerve-fiber arborizations, and by regenerative branching to muscle fibers thus denervated either from terminal nerve fibers, or from the motor end plate itself. This latter '"ultraterminal branching" is said to be a specific feature of diabetic neu- ropathy (2). When these branches reach muscle fibers they form new end plates. In early dia- betes the new branches are well-formed and healthy, but in more advanced cases they are thin, finely beaded, with few and poorly developed end plates. Thus in advanced diabetes there may be severe degeneration of the terminal neuromuscular apparatus accompanied clinically by peripheral muscle wasting. Skeletal Muscle In coincidence with the early changes in the motor end plates, muscular abnormalities are present at the onset of juvenile-type diabetes. These vary from individual slender fibers to areas of clear-cut neurogenic atrophy (68). Whether or not such changes invariably remain after treatment, or occur early in milder diabetes, is unknown. They usually are not clinically mani- fest and do not show a strong predisposition to advance, presumably in part because of continuing regeneration of end plates. Electron microscopy of muscle in early diabetes has not been done. Electron microscopy in patients with clinical neuropathy reveals wide separation of myofibrils, abnormal glycogen deposi- tion, swelling of mitochondria, dilation of the sarcoplastic reticulum, and lysosome formation (2). The extent to which these changes are neurogenic and nonspecific, and to which they are diabetes specific, needs to be established. Diabetic amyotrophy is a rare syndrome characterized by rapid onset of a profound proximal motor deficit and muscle wasting and a relatively good prognosis for recovery. This is in con- trast to the usual late development of mild, chronic, peripheral muscle wasting in diabetic neuropathy. There is a general present belief that the muscle dystrophy is of neurogenic rather than myogenic origin (10). This belief needs to be reviewed. It is noteworthy that the mild disseminated muscle atrophy seen histologically may far from reflect the severity of the weakness, and that at the time of return of function the same biopsy changes may persist (57). It could be suspected, therefore, that these changes are nonspecific, preceded the amyotrophy, and that actually nothing is known about the morphologic counterparts of this remarkable clinical condition. PATHOPHYSIOLOGY Nerve Conduction Velocity The most apparent electrophysiologic characteristic of diabetic peripheral neuropathy is an early slowing of nerve conduction velocity by approximately 30% from normal values, followed by further slowing as the disease progresses (28,52,56,81). Reduced conduction velocity results Diabetic Peripheral Neuropathy 247 from degenerative changes in the myelin sheath, with a functional effect of sorbitol accumulation perhaps also being implicated (40). Diabetic neuropathy shares this characteristic with other demyelinating disorders such as diphtheritic neuropathy and subacute combined degeneration, in contrast to axonal disorders such as alcoholic and uremic neuropathy in which there is little or no early change in conduction velocity. The reason for the lessened conduction velocity in diabetes and other demyelinating disorders is probably a change in the electrical properties of the myelin sheath (21,54,79). Conduction velocity in myelinated nerves is proportional to fiber resistance. Resistance is proportional to myelin thickness, myelin lipids being the principal determinant. Thus resistance along the internodes is 10° times that of the nodes. Internodal resistance is diminished within days of the the onset of diabetes (30). The internodes act as passive core conductors, the membrane action potential being regenerated at each node (61). Conduction velocity depends principally upon the time taken at successive nodes for the current to depolarize the axon membrane to the critical level necessary for triggering an action potential. As the insulating property of the myelin diminishes, it is likely that increased leakage of current along the internodes reduces the cur- rent density at the nodes, delaying their excitation. The early drop in internodal resistance in diabetic neuropathy may be due in part to changes in myelin fatty-acid chain length, saturation or branching. It probably also is related to the early paranodal separation of myelin lamellae (3), perhaps due to interlamellar fluid accumulation (81), a process in which sorbitol concentra- tion might play a part. Impaired sensory perception is a functional expression of changes in conduction velocity. It has been demonstrated shortly after the clinical onset of diabetes in all areas and modalities that have been tested. Perception thresholds for light touch, two-point discrimination and vi- bration in the hands and feet, and for corneal touch, visual flicker, auditory flutter, and taste, are all diminished early in diabetes (12,44 ,74). The diminution in conduction velocity occurs unequally amongst nerve fibers, so that not only the rate, but the amplitude and shape of sensory- nerve action potentials become abnormal (52). Temporal dispersal of the action potential likely blurs acuity of definition at the cerebral cortex. Electromyography Whether or not there is a corresponding functional expression in the motor system of the early changes in conduction velocity and in the structure of motor end plates is less clear. In one study (16) electromyograms were normal in early diabetes. In another study (52), however, fibrillation potentials, motor unit loss, and abnormalities of the shape and duration of muscle action potentials had developed in some patients within a few months of onset, although these changes were more common in case of longer duration. Abnormalities in muscular performance accompanying these early electrophysiologic findings, analogous to the early deficits in sensory perception, have not been described. This may be because insufficient care and sophistication have been employed to demonstrate them. An attempt (49) to explore the onset of electromyographic abnormalities in experimental dia- betes seems to have been unsuccessful, perhaps because of the added factor of nutritional changes due to glycosuria. The morphologic and electrophysiologic changes did not correspond to those seen in diabetes in man. A systematic electromyographic study in diabetic amyotrophy seems not to have been done. The 248 Diabetes Mellitus findings have suggested a primary nerve disease (13, 52), but as with the morphologic changes, may be mild in relation to clinical severity. One thus wonders again whether they were related to the myopathy, or whether they preexisted it. Amyotrophy may be a functional muscle disorder with pre- sently unknown morphologic, electrical and biochemical characteristics. Response to Ischemia The response of nerve function to ischemia is abnormal in diabetes. Vibratory perception has been studied most fully (78,80). Following inflation of a limb cuff in normal subjects the vibra- tory perception threshold remains normal for 10 to 20 minutes, and then swiftly rises until vibra- tion can no longer be felt. In diabetic subjects the threshold during ischemia remains normal for 30 minutes or longer. The response to ischemia can be restored to normal by careful insulin therapy. More recently this phenomenon in diabetes has been shown to include motor-nerve conduction velocity, sensory-nerve action potentials, and touch perception thresholds (45,80). Pain- and heat-perception thresholds do not diminish during ischemia even in health. The specificity of the phenomenon has not been defined. It occurs in some healthy aged individuals (23) and in uremia (15), and is said to occur in pernicious anemia (80). Preservation of function during anoxia occurs in isolated diabetic nerve, so that the phenomenon is a property of the nerve fibers themselves rather than of the metabolites released by surrounding ischemic tissues (75). Its cause is unknown. Suggestions are a greater ability to maintain energy synthesis by anaerobic glycolysis, increased potassium diffusion away from nerve through a damaged Schwann-cell basement membrane, and facilitation of nerve conduction by focal areas of demyelination acting as points of depolarization. Autonomic Nervous System Systematic neurophysiologic studies are lacking. Heightened esophageal responsiveness to stretch (50), impaired papillary responsiveness to light (39), and cholecystic distension (42) in early diabetes suggest a generalized disorder of autonomic function well before the appearance of clinical neuropathy. BIOCHEMISTRY Structural Analysis Peripheral myelin is an elaboration of Schwann-cell plasma membrane (5). Its biochemical analysis is at a relatively primitive stage. It consists of alternating protein and lipid layers. Compared to central myelin it has fewer sedimentary subfractions, a differing amino-acid composi- tion, and differing proportions of cholesterol, triglyceride, galactolipids, phospholipids and polyunsaturated fatty acids (20,65,66,85). The sedimentary and electrophoretic subfractions of peripheral myelin in diabetes differ from normal (65). It is not known whether these structural differences are synthetic or degrada- tive in nature. The breakdown of myelin in disease has been studied in several experimental disorders but not in diabetic neuropathy (47). The early lamellar splitting in myelin breakdown is accompanied by a progressive rise in several proteinases originating from Schwann-cell lysosomes and surface membranes. Metabolic Abnormalities of Diabetic Nerve and Effects of Insulin Glucose uptake, glucose oxidation, and glucose-derived lipogenesis in peripheral nerve are insulin responsive (36, 37). It is curious that glucose uptake is not insulin dependent, being actually higher in diabetic than in normal nerve. The action of insulin or glucose uptake is a Diabetic Peripheral Neuropathy 249 sterospecific cell-membrane phenomenon. Insulin also influences the intracellular disposition of glucose. Normal glycogen stores are absent and lipogenesis is depressed in diabetic nerve, while lipogenesis is increased in normal nerve even in a glucose-free medium. Sorbitol and fructose accumulate in diabetic nerve in the presence of depressed insulin-dependent pathways. More com- prehensive, structure-specific (axon vs satellite cell) studies are needed to identify the rela- tionship of these various observations to the morphologic and functional changes present in diabetic nerve. Insulin markedly enhances the total incorporation of 14C-acetate into normal-nerve lipid, and alters its fractional incorporation into lipid components. Diabetes likewise alters these incor- porations (29,37,66). Details within and amongst studies are confusing and do not lead to a comprehensible picture, perhaps because of an influence of cachexia as well as of diabetes, because the observations are not structure specific, and because pool sizes and specific activities were not measured. The Sorbitol Pathway The permeability of peripheral nerve to glucose is not insulin dependent, and hence its intra- cellular glucose concentrations is at the extracellular level. Its intracellular glucose disposal to glycogen and to lipids is insulin dependent, whereas disposal via the sorbitol pathway [glucose (aldose reductase) — sorbitol (sorbitol dehydrogenase) —— fructose] is not, so that at any given blood glucose level diabetic nerve has an even greater tendency to form sorbitol and fructose than does normal nerve (40, 41). These sugars traverse cell membranes poorly, and are metabolized only slowly, so that once formed they are trapped intracellulary. Nerve from diabetic patients contains sufficient sorbitol and fructose to have osmotic significance (10-80 mM/g). Furthermore, aldose reductase is located within the Schwann cells and sorbitol dehydrogenase mainly within the axon, so that local concentrations may be even higher than is apparent from whole tissue levels. In diabetic rat nerve the defect in conduction velocity is concomitant with and proportional to the elevation of sorbitol and fructose (40, 41). Control of the blood glucose level with insulin restores both abnormalities to normal. The problem of establishing causality between them was approached by using galactose-fed rats. Galacticol, the analogue of sorbitol, accumulates in the nerves of these rats, coincident with which there is an increase in water content and a decrease in conduction velocity. Removal of galactose from the diet or administration of an aldose-reductase inhibitor corrects both abnormalities. The necessary critical experiments to establish the analo- gous causal link in diabetes have not been performed. It would be desirable that such observations be coupled with studies of myelin ultrastructure, composition, and electrical resistance. Other Biochemical Properties of Nerve Other properties of peripheral nerve could merit consideration in relation to the etiology or pathogenesis of diabetic peripheral neuropathy. Nerve Growth Factor: Differentiation, growth and maintenance of the sympathetic nervous system in animals is dependent upon a specific protein having a structure and metabolic effect similar to proinsulin and to insulin respectively (38,55). The significance, if any, of these observations to health and disease has not been determined. Axoplasmic Transport in Nerve: A slow and a fast transport mechanism carry protoplasmic materials from the cell body out into the axon (53,63,64). Slow transport is by means of an unexplained convection of whole axoplasm. Fast transport is by a sliding-filament mechanism. It is dependent 250 Diabetes Mellitus upon ATP generated by local oxidative metabolism, probably within the Schwann cell. The influence of diabetes upon these transport mechanisms has not been examined. Role of Cyclic AMP in Nerve Impulse Transmission: Cyclic AMP has a role in the release of neuro- transmitters from synapses and from motor end plates which is analogous to its regulation of the secretion of subcellular vesicles from endocrine glands (11,43,60,71). Study of the mechanism is incomplete. Cyclic AMP promotes glycolytic metabolism in nerve endings which may trigger the calcium-dependent release mechanism, or it may have a direct effect on calcium distribution. The effect of diabetes on this system has not been explored. TREATMENT The clinical literature and textbooks on diabetes mellitus convey an impression that the occurrence and progress of significant diabetic neuropathy are unrelated to the quality of dia- betes management. This is untrue. The greatest need in the study of diabetic neuropathy now is a careful prospective clinical study which readily would dramatize this palpable fact. The results of such a study, and of similar ones related to the peripheral angiopathic complications of dia- betes, could crystallize agreement amongst physicians that the principles and the details of currently-available diabetes treatment, primitive though it is, are effective, and could be used to motivate physicians and patients to learn and to apply them. Such a new era in diabetes management, pending the development of new insulin-delivery systems, is available now. The preventibility of diabetes complications is in fact most clear for those disorders in which a satellite cell is the target for damage: the axon-supporting Schwann cell in neuropathy, and the capillary-supporting pericyte (17) in retinopathy. Quantitative studies (19) of the rela- tionship of the severity of damage to the quality of treatment support one's own experience that severe, crippling neuropathy and retinopathy occur only in those individuals whose treatment has been negligent. The principal present management of diabetic neuropathy is its prevention, or its arrest if it already is present. An acute reversal to normal by insulin of the early functional abnormali- ties in nerve conduction velocity and in the axon response to ischemia is demonstrable both in experimental and in clinical diabetes (40,45,46,80). Chronic diabetic neuropathy, however, tends to be intractable. It is not impervious to treatment effects, particularly those of meticu- lous, insulin-based diabetes control. However reports of its improvement with any regimen need to be evaluated against the fact that its symptoms are characterized by spontaneous remissions and exacerbations (59). Thus, since attention is likely to be sought by a patient during an exacer- bation, improvement is likely to occur contemporaneously with treatment. In particular, sympto- matic relief with the neurodepressors diphenylhydantoin (32) and carbamazepine (72), and the lipid-lowering agent chlorophenoxyisobutyrate (6,26) need confirmation. Thiamine chloride (18) and cyanocobalamin (4) also often are used, although firm evidence that they change the course of the disease likewise is lacking. It is interesting that latent vitamin B12 deficiency coexists with diabetes, possibly on an autoimmune basis, in about 3 percent of cases (84) and in about 9 percent of cases with peripheral neuropathy (51). The neuropathic signs in many of these indi- viduals, originally attributed to diabetes, were responsive to vitamin B12. This observation is of therapeutic relevance in itself, but also it leads one to wonder whether in diabetic nerve the prevailing disorders of fatty acid and of pyruvate (62,83) metabolism might somehow lower the Diabetic Peripheral Neuropathy 251 thresholds for the manifestations of cofactor depletion, and whether the administration of cyano- cobalamin and of thiamin may in fact have a useful therapeutic place on this basis. One retains a clinical impression that they are helpful, and investigations of this point seem appropriate. At the level of applied research, the need in diabetic neuropathy as in other diabetes com- plications, is the development of transplantable or electronic insulin- and possibly glucagon- delivery systems which have sensing and input characteristics more closely resembling those which naturally occur. At the level of basic research, the greatest need in diabetic neuropathy is definition of which enzyme reaction(s) in the Schwann cell is so critically insulin dependent that morphologic and functional disturbances develop in it from the moment of onset of diabetes. A comparative study of central and of peripheral satellite cells might assist in this search. SUMMARY Neuropathy is the earliest and the most universal structural complication of diabetes, co- existing in fact with the disease from its onset. It probably is caused by a specific biochemical lesion within the axon-satellite cell of Schwann, and all of its manifestations probably are secondary to this primary defect. It is the complication of diabetes most supressible even with present modes of management. Specific definition of its etiology and pathogenesis and development of improved modes of management leading to full preventibility are probably within the grasp of present technology. REFERENCES 1. Appenzeller, O, and EP Richardson Jr 1966. The sympathetic chain in patients with diabetic and alcoholic polyneuropathy. Neurology 16:1205-1209. 2. Awad, EA 1970. Motor-point biopsies in diabetic neuropathy. Arch Phys Med 51:418-422. 3. Babel, J, A Bischoff, and H Spoendlin 1970. Ultrastructure of the peripheral nervous system and sense organs. Atlas of normal and pathologic anatomy. Saint Louis, Missouri. CV Mosby Co. 4. Becker, B, GD Maengwyn-Davies, D Rosen, JS Friedenwald, and FC Winter 1954. The adrenal cortex and B-Vitamins in diabetic retinopathy. Diabetes 3:175-187. 5. Ben Geren, B 1954. The formation from the Schwann cell surface of myelin in the peripheral nerves of chick embryos. Exp Cell Res 7:558-562. 6. Berenyi, MR, B Straus, and OE Miglietta 1971. Treatment of diabetic neuropathy with clofi- brate. J Amer Geriat Soc 19:763-772. 7. Bischoff, A 1968. Diabetic neuropathy. Morbid anatomy, patho-physiology and pathogenesis based on electron-microscope findings. Germ Med Mth 13:214-218. 8. Bischoff, A 1973. Ultrastructural pathology of peripheral nervous system in early diabetes. In advances in metabolic disorders. Supplement 2. Vascular and neurological changes in early diabetes. RA Camerini-Davolos and HS Cole, editors. Academic Press, New York pp 441-449. 9. Budzilovich, GN 1970. Diabetic neuropathy complex. Virchows Arch Abt A Path Anat 350:105-122. 10. Casey, EB, and MJG Harrison 1972. Diabetic amyotrophy: A follow-up study. Brit Med J 1:656-659. 252 Diabetes Mellitus 11, 12. 13, 14. 15, 16. 17. 18. 19. 20. 21. 22. 23. 24, 25. 26. 27. 28. 29. 30. 31. Cedar, H, ER Kandel, and JH Schwartz 1972. Cyclic adenosine monophosphate in the nervous system of aplysia californica I. Increased synthesis in response to synaptic simulation. J Gen Physiol 60:558-569. Chochinov, RH, GLE Ullyot, and JA Moorhouse 1972. Sensory perception thresholds in patients with juvenile diabetes and their close relatives. New Engl J Med 286:1233-1237. Chopra, JA and LJ Hurwitz 1968. Femoral nerve conduction in diabetes and chronic occulsive vascular disease. J Neurol Neurosurg Psychiat 31:28-33. Chopra, JS, LJ Hurwitz, and DAD Montgomery 1969. The pathogenesis of sural nerve changes in diabetes mellitus. Brain 92:391-418. Christensen, NJ, and H Orskov 1969. Vibratory perception during ischaemia in uraemic patients and in subjects with mild carbohydrate intolerance. J Neurol Neurosurg Psychiat 32:519-524. Coers, C, and J Hildebrand 1965. Latent neuropathy in diabetes and alcoholism. Neurology 15:19-38. Cogan, DG, D Toussaint, and T Kuwabara 1961. Retinal Vascular Patterns. IV Diabetic retino- pathy. Arch Opth 66:366-378. Collens, WS, AM Rabiner, JD Zilinsky, LC Boas, and JJ Greenwald 1950. The treatment of peri- pheral neuropathy in diabetes mellitus. Amer J Med Sci 219:482-487. Collyer, RT, and BE Hazlett 1961. Retinopathy and neuropathy in one hundred growth-onset diabetic patients. Can Med Assoc -J 85:1328-1334. Davison, AN, and A Peters 1970. Myelination. Springfield, Illinois. Charles C Thomas. Davson, H 1970. Textbook of General Physiology. 4th edition. London, J § A Churchill Ltd. Dickens, WN, RK Winkelmann, and DW Mulder 1963. Cholinesterase demonstration of dermal nerve endings in patients with impaired sensation. Neurology 13:91-100. Digiesi V, A Malavasi, and D Giraudi 1969. Effetto dell'ischemia acute sulle sensibilita vibratoria, termica, tattile e dolorifica in soggetti di etd senile. Rass Neur Veg 23:71-82. Dolman, CL 1967. The pathology and pathogenesis of diabetic neuropathy. Bull NY Acad Med 43:773-783. Dotevall, G, SE Fagerberg, L Langer, and A Walan 1972. Vagal fucntion in patients with diabetic neuropathy. Acta Med Scand 191:21-24, Duncan, GG, FA Elliott, TG Duncan, and J Schatanoff 1968. Some clinical potentials of chlorophenoxyisobultyrate (chofibrate) therapy (Hyperlipidemia-Angina-Pectoris-blood sludging- diabetic neuropathy). Metabolism 17:457-473, Dyck, PJ, RK Winkelmann, and CF Bolton 1966. Quantitation in Meissner's corpuscles in hereditary neurologic disorders. Charcot-Marie-Tooth disease, Roussy-Levy snydrome, Dejerine- Sottas disease, hereditary sensory neuropathy, spino-cerebellar degenerations, and hereditary spastic paraplegia. Neurology 16:10-17. Eliasson, SG 1965. Nerve conduction changes in experimental diabetes. J Clin Invest 43:2353-2358. Eliasson, SG 1966. Lipid synthesis in peripheral nerve from alloxan diabetic rats. Lipids 1:237-240. Eliasson, SG 1969. Properties of isolated nerve fibers from alloxanized rats. J Neurol Neurosurg Psychiat 32:525-529. Ellenberg, M 1960. Diabetic neuropathy. A consideration of factors of onset. Ann Int Med 52:1067-1075. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42, 43. 44. 45. 46. 47. 48. 49. 50. 51. Diabetic Peripheral Neuropathy 253 Ellenberg, M 1968. Treatment of diabetic neuropathy with diphenylhydatoin. New York J Med 68:2653-2655. Ellenberg, M 1973. Current status of diabetic neuropathy. Metabolism 22:657-662. Faerman, I, D Fox, MN Jadzinsky, L Glocer, and JB Cibeira 1973a. Neurological findings in chemical diabetes. In advances in metabolic disorders. Supplement 2. Vascular and neuro- logical changes in early diabetes. RA Camerini-Davolos, and HS Cole, editors. Academic Press. New York. pp 451-457. Faerman, I, L Glocer, D Celener, M. Jadzinsky, D Fox, M Maler, and E Alvarez 1973b. Autonomic nervous system and diabetes. Histological and histochemical study of the autonomic nerve fibers of the utinary bladder in diabetic patients. Diabetes 22:225-237. Field, RA, and LC Adams 1964. Insulin response of peripheral nerve. I Effects on glucose metabolism and permeability. Medicine 43:275-279. Field, RA, and LC Adams 1965. Insulin response of peripheral nerve. II Effects on lipid metabolism. Biochim Biophys Acta 106:474-479. Frazier, WA, RH Angeletti, and RA Bradshaw 1972. Nerve growth factor and insulin. Structural similarities indicate an evolutionary relationship reflected by physiological action. Science 176:482-488. Friedman, SA, R Feinberg, E Podolak, and RHS Bedell 1967. Papillary abnormalities in diabe- tic neuropathy. A preliminary study. Ann Int Med 67:977-983. Gabbay, KH 1973a. The sorbital pathway and the complications of diabetes. New Engl J Med 288:831-836. Gabby, KH 1973b. Role of sorbitol pathway in neuropathy. In advances in metabolic disorders. Supplement 2. Vascular and Neurological Changes in Early Diabetes. RA Camerini-Davolos, and HS Cole, editors. Academic Press. New York. pp 417-424, Gitelson, S, D Oppenhein, A Schwartz 1969. Size of the gall bladder in patients with diabetes mellitus. Diabetes 18:493-498. Goldberg, AL, and JJ Singer 1969. Evidence for a role of cyclic AMP in neuromuscular trans- mission. Proc Natl Acad Sci USA 64:134-141. Gregersen, G 1968a. Vibratory perception threshold and motor conduction velocity in diabetes and nondiabetics. Acta Med Scand 183:61-65. Gregersen, G 1968b. A study of the peripheral nerves in diabetic subjects during ischaemia. J Neurol Neurosurg Psychiat 31:175-181. Gregersen, G 1968c. Variations in motor conduction velocity produced by acute changes of the metabolic state in diabetic patients. Diabetologia 4:273-277. Hallpike, JF, and CWM Adams 1969. Proteolysis and myelin breakdown: a review of recent histo- chemical and biochemical studies. Histochem J 1:559-579. Hensley, GT and KH Soergel 1968. Neuropathologic findings in diabetic diarrhea. Arch Path 85:587-597. Hildebrand, J, A Joffroy, G Graff, and C Coers 1968. Neuromuscular changes with alloxan hyperglycemia. Arch Neurol 18:633-641. Horgan, J, and JS Doyle 1969. Manometric oesophageal motility studies in diabetes without neuropathy. Irish J Med Sci 2:475-480. Khan, MA, GS Wakefield, and DW Pugh 1969. Vitamin-B12 deficiency and diabetic neuropathy. Lancet 2:768-770. 254 Diabetes Mellitus 52. Lamontagne, A, and F Buchthal 1970. Electrophysiological studies in diabetic neuropathy. J Neurol Neurosurg Psychiat 33:442-452, 53. Lasek, RJ 1970. Protein transport in neurones. International Review of Neurobiology vol 13. CC Pfeiffer, and JR Smythies, editors. pp 289-324, 54. Leslie, RB, D Chapman, and CJ Hart 1967. Physical studies of phospholipids. VII The D.C electrical conductivity properties of some membrane phospholipids. Biochim Biophys Acta 135:797-811. 55. Liuzzi, A and FH Foppen 1968. Sterol-like compound from sensory ganglia. Effect of a nerve growth factor and insulin on its biosynthesis. Biochem J 107:191-196. 56. Locke, S 1967. Axons, Schwann cells and diabetic neuropathy. Bull NY Acad Med 43:784-791. 57. Locke, S, DG Lawrence, and MA Legg 1963. Diabetic amytrophy. Amer J Med 34:775-785. 58. Majno, G, and ML Karnovsky 1958. A biochemical and morphologic study of myelination and demyelination. 1. Lipid biosynthesis in vitro by normal nervous tissue. J Exp Med 107:475-496. 59. Mayne, N 1968. The short-term prognosis in diabetic neuropathy. Diabetes 17:270-273. 60. McAfee, DA, M Schorderet, and P Greengard 1971. Adenosine 3', 5'-monophosphate in nervous tissue: Increase associated with synaptic transmission. Science 171:1151-1158. 61. McDonald, WI 1963. The effects of experimental demyelination on conduction in peripheral nerve: A histological and electrophysiological study II Electrophysiological observations. Brain 86:501-524. 62. Moorhouse, JA 1964. Pyruvate-tolerance tests in healthy and diabetic subjects. Lancet 1:689-693. 63. Ochs, S 1971. Characteristics and a model for fast axoplasmic transport in nerve. J of Neurobiol 2:331-345. 64. Ochs, S, and J Johnson 1969. Fast and slow phases of axoplasmic flow in ventral root nerve fibres. J of Neurochem 16:845-853, 65. Palo, J, H Savolainen, and M Haltia 1972. Proteins of peripheral nerve myelin in diabetic neuropathy. J Neurol Sci 16:193-199. 66. Pratt, JH, JF Berry, B Kaye, and FC Goetz 1969. Lipid class and fatty acid composition of rat brain and sciatic nerve in alloxan diabetes. Diabetes 18:556-561. 67. Reske-Nielsen, E, and K Lundbaek 1968. Pathological changes in the central and peripheral nervous system of young long-term diabetics II The spinal cord and peripheral nerves. Dia- betologia 4:34-43. 68. Reske-Nielsen, E, G Gregersen, A Harmsen, and R Lundboek 1970a. Morphological abnormalities of the terminal neuromuscular apparatus in recent juvenile diabetes. Diabetologia 6:104-109. 69. Reske-Nielsen, E, K Lundboek, G Gregersen, and A Harmsen 1970b. Pathological changes in the central and peripheral nervous system of young long-term diabetics. The terminal neuro- muscular apparatus. Diabetologia 6:98-103. 70. Reske-Nielsen, E, R Lundbaek, and OJ Rafaelsen 1965. Pathological changes in the central and peripheral nervous system of young long-term diabetics. Diabetologia 1:233-241. 71. Rubin, RP 1969. The metabolic requirements for catecholamine release from the adrenal medulla. J Physiol 202:197-209. 72. Rull, JA, R Quibrera, H Gonzalez-Millan, and OL Castafleda 1969. Symptomatic treatment of peripheral diabetic neuropathy with carbamazepine (Tegretol): Double blind crossover trial. Diabetologia 5:215-218. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82, 83. 84. 85. 86. Diabetic Peripheral Neuropathy 255 Schmitt, FO 1958. Axon-satellite cell relationships in peripheral nerve fibers. Exp Cell Res Suppl 5:33-57. Schwartz, DE 1974. Corneal sensitivity in diabetes. Arch Ophthal 91:174-178. Seneviratne, KN, and OA Peiris 1968. The effect of ischaemia on the excitability of sensory nerves in diabetes mellitus. J Neurol Neurosurg Psychiat 31:348-353. Seneviratne, KN, and OA Peiris 1969. The effects of hypoxia on the excitability of the isolated peripheral nerves of alloxan-diabetic rats. J Neurol Neurosurg Psychiat 32:462-469. Singer, M, and MM Salpeter 1966. The transport of 3H-1-histidine through the schwann and myelin sheath into the axon, including a reevaluation of myelin function. J Morph 120:281-316. Steiness, I 1961. Influence of diabetic status on vibratory perception during ischaemia. Acta Med Scand 170:319-338. Tasaki, I 1955. New measurements of the capacity of the resistance of the myelin sheath and the nodal membrane of the isolated frog nerve fiber. Amer J Physiol 181:639-650. Terkildsen, AB and NJ Christensen 1971. Reversible nervous abnormalities in juvenile diabetics and recently diagnosed diabetics. Diabetologia 7:113-117. Thomas, KP 1971. The morpholgical basis for alterations in nerve conduction in peripheral neuropathy: Proc Roy Soc Med 64:295-298. Thomas, PK, and RG Lascelles 1966. The pathology of diabetic neuropathy. Quart J Med 140:489-509. Thompson, RHS, WJIH Butterfield, and IK Fry 1960. Pyruvate metabolism in diabetic neuropathy. Proc Roy Soc Med 53:143-146. Ungar, B, AE Stocks, FIR Martin, S Whittingham, and IR MacKay 1968. Intrinsic-factor antibody, parietal-cell antibody, and latent pernicious anaemia in diabetes mellitus. Lancet 2:415-417. Wolfgram, F, and K Kotorii 1968. ' The composition of the myelin proteins of the peripheral nervous system. J of Neurochem 15:1291-1295. Zitomer, BR, HF Gramm, and GP Kozak 1968. Gastric neuropathy in diabetes mellitus: Clinical and radiologic observations. Metabolism 17:199-211. 19 EFFECT OF DIABETES MELLITUS ON FETAL GROWTH AND DEVELOPMENT! Daniel H. Mintz and Ronald A. Chez BACKGROUND Prior to the introduction of exongenous insulin therapy, few women with diabetes mellitus were able to become pregnant. Those who did faced a further threat to their existence, since pregnancy superimposed unique and unmet metabolic requirements on an already compromised nutritional state. Without insulin therapy, ketoacidosis was an inevitable terminating event for both mother and her unborn child. Following the introduction of insulin therapy, more women reached the childbearing ages; fer- tility was unimpaired, and the spontaneous abortion rate reverted to that encountered in nondia- betic women. The dramatic reduction in maternal mortality then served to uncover a relentless and unremitting perinatal morbidity and mortality which has characterized pregnancies complicated by diabetes mellitus in the last five decades. Diabetes mellitus is recognized increasingly as an accompaniment of pregancy. It is presently conservatively estimated that one out of each 100 pregnancies is complicated by diabetes (61). -In the decades ahead this prevalence rate will undoubtedly increase, ultimately placing diabetes mellitus as a leading medical complication of pregnancy. Considerable progress has been made in the last two decades in providing a clearer under- standing of substrate and hormone dependence and independence in the placental fetal maternal unit. This report will briefly summarize the current extent of our information regarding fetal maternal carbohydrate metabolism in normal gestation. Limited insights which are now available, concerning the pathophysiologic alterations imposed by diabetes mellitus, will also be reviewed. Several reports can be consulted for more detailed treatment (9,27,38.46.53). MATERNAL METABOLISM As normal gestation progresses a lowering in the concentration of blood glucose occurs in the fasted state (22). These changes are accompanied by an elevation in plasma free fatty acids and ketone bodies and a decrease in plasma amino acids, particularly gluconeogenic precursors. Freinkel (21) has emphasized that these changes mimic the alterations observed in prolonged starvation in the nongravid state. He postulated that pregnancy was manifestly a state of accelerated starva- tion, with the conceptus serving as an obligate parasite. Pregnancy is also accompanied by an elevation in fasting immunoreactive insulin, enhanced glucose mediated beta cell responsiveness despite diminished tolerance to orally administered glucose, an increase in the fractional turnover for maternal insulin, decreased peripheral insulin sensitivity, diminished growth hormone responses to provocative stimuli in late gestation and a progressive plasma elevation of human placental lactogen (HPL), a peptide secreted by the placenta into the maternal circulation at a rate propor- tional to the trophoblastic mass. Placental lactogen appears to play a direct role in some, but lEditorial comment: Various other aspects of the problem of diabetes in pregnancy are considered in chapters 2 and 8. 256 Effects on Fetal Growth and Development 257 clearly not all, of the changes described above (6, 32). Moreover, it is unlikely that these multiple alterations will ever submit to a solitary signal that then initiates a sequential physiologic cascade. For example, the enhanced hepatic fractional turnover rate of gluconeogenic precursors may be related to a placental factor (42) other than placental lactogen; the increased beta cell responses to insulinogenic stimuli may relate to steroid hormones derived from the pla- centa (5) or augmentation in insular blood flow (43), and the alteration in the pituitary growth hormone releasing mechanism may be due to time-dependent opposing effects of estrogens and HPL derived from the placenta (48). In this emerging hypothesis the conceptus, more specifically its placental secretory products, is the critical regulator of metabolic events in the maternal com- partment; its transport function, to the contrary, may play only a passive role in these changes. The net result of these alterations in the pregnant female is to preserve glucose homeostasis. The enhanced insulin secretory capacity of the maternal pancreatic beta cell is critical to the adaptive response. Thus in the normal pregnant subject, food intake is associated with an aug- mented insulin response and only minimal and transient elevations in maternal plasma glucose level. The fetus, in turn, is supplied with a more or less constant nonperturbating source of nutrients from the mother across the placenta, despite the exigencies of maternal food consumption. To the extent that the maternal islet fails to compensate for the metabolic stress, the diabetogenic effect of pregnancy is manifested as: (a) gestational diabetes, (b) asymptomatic. chemical aiabetes, and (c) an intensification of the insulin dependent state in a previously recognized diabetic pa- tient. When the efficiency of the maternal beta cell is reduced, the fetus must then develop its own adaptive mechanism to offset the effects of a fluctuating source of nutrition. FETAL METABOLISM A. Development and function of the normal human endocrine pancreas. Alpha and beta cells can be identified in the human pancreas at 50 days of post-conception life (55). Our present knowledge about the morphogenesis and development of the organ has been summarized elsewhere (33,44). It would appear that a positive correlation exists between the number of alpha cells and beta cells and the amount of extractable immunoreactive glucogan and insulin; the percentage of either cell per total pancreatic cell mass is higher in the fetus than in the adult as is the insulin and glucagon content per gram weight of tissue. Furthermore the alpha:beta cell ratio of five in early pregnancy is one at birth and then approaches adult levels of 0.2 to 0.1 in childhood. The precise role(s) served by the relative excess of alpha cells to beta cells in early life is unknown. Although the earliest appearance of glucagon in fetal plasma has not been precisely deter- mined, it does appear that fetal pancreatic and plasma glucagon (3) may be present at a stage of gestation that precedes the development of glucagon receptors (13) or enzymatic pathways responsive to the hormone (28). Whether fetal pancreatic glucagon in early gestation has a role in beta cell ontogenesis has yet to be determined. Moreover, it has not yet been determined which stimuli, if any, affect changes in fetal plasma concentrations. In early reports, neither neonatal hypergly- cemia nor maternal diabetes mellitus has been associated with changes in the plasma glucagon levels of the conceptus (7, 41). It does appear that perinatal hypoxia (29), intrauterine growth retar- dation (7) and neonatal hypoglycemia may influence plasma glucagon concentrations (7, 41). Insulin containing granules can be identified by 60 days of gestation (55). The presence of insulin in fetal plasma has been observed at 84 days of gestation (1). The basal level of fetal 258 Diabetes Mellitus insulin, once achieved, remains relatively constant in a range similar to those observed in adult life. The nature of the driving force to maintain this basal level has not been determined. Amino acids, glucagon, and theophylline; when administered intravascularly are associated with a rise in fetal or neonatal plasma insulin (26, 36). Therefore essential components of the insulin releasing mechanism are present in the fetal and neonatal pancreatic beta cell. A recent review (8) analyzed the human fetal and newborn pancreatic beta cell responsiveness to glucose. In almost all studies, the administration of glucose is associated with either the absence of or the delayed release of insulin. The mechanism responsible for the attenuated fetal response to a glycemic challenge, and the factor(s) influencing the gradual shift to an adult beta cell response pattern in early neona- tal life are unknown. Whether a glucoreceptor or the intrinsic metabolic pathways of the beta cell required for glucose mediated insulin release are absent, inhibited, or not fully developed in the fetus is not clear. The absence of glucose mediated insulin responses in the fetus could be likened to islet cell function of the juvenile diabetic, whereas the delayed insulin responsiveness of the 1- to 3-day-old neonate resembles the pattern of release observed in maturity onset diabetes. Further in vitro study of islets from this transitional period in islet cell function may offer important insights into the mechanism(s) of control of the alpha and beta cells in both individuals and diabetic patients. B. Effect of diabetes mellitus on the fetal endocrine pancreas The transfer of glucose across the placenta is secondary to the driving force of the concen- tration gradient between mother and fetus. A direct relationship exists between maternal, fetal, and amniotic fluid concentrations of glucose (11). Thus in the patient with either fasting or postprandial hyperglycemia, the maternal glucose increment is proportionately reflected in the fetal blood compartment. Placental saturation kinetics for maternal to fetal glucose transfer have been postulated in man (4), but its presence has not been confirmed in studies in the subhuman primate (10). Pedersen (52) proposed that prolonged exposure of the fetus in a diabetic pregnancy to hyper- glycemia stimulated the fetal pancreas to produce excessive insulin. Fetal hyperglycemia and hyperinsulinemia together promoted glucose uptake ifi fetal adipose tissue leading to the obesity (20) characteristic of infants of diabetic mothers. The complication of neonatal hypoglycemia (14) could then be related to the persistence of pancreatic beta cell hyperresponsiveness into the neonatal period. The absence of the embryopathy characteristic of infants of diabetic mothers when maternal hyperglycemia was moderately controlled provided apparent confirmation of this patho- physiologic sequence (34,50,52). Moreover, increased fetal newborn basal insulin levels and the plasma insulin responses to a glycemic stimulus in newborns of pregnancies associated with diabetes mellitus were also in accord with this hypothesis (34,45,49). There are, however, certain persuasive observations which suggest that fetal hyperglycemia/ hyperinsulinemia may not be a sufficient reason for the embryopathy in infants of a diabetic mother. First, neonatal hyperinsulinemia has not been a consistent finding in all studies of infants of hyperglycemic mothers (35). Second, the structural integrity of the fetal neurohypo- physis is required for the emergence of fetal hyperinsulinemia even in the presence of fetal hyper- glycemia (15). Third, the classical appearances of these fetuses are not noticeable before a gestational age of about 34 weeks in spite of the presence of maternal/fetal hyperglycemia since Effects on Fetal Growth and Development 259 conception (19). Fourth, fetal hyperinsulinemia also accompanies erythroblastosis (17) and other fetal hemolytic disorders (60), but fetal obesity does not. Last, in humans (59) and in induced hyperglycemia in subhuman primate pregnancies (47), fetal basal hyperinsulinemia can occur in normoglycemic fetuses. These observations indicate factors other than fetal hyperglycemia may influence the fetal endocrine pancreas. The role and interrelationships of other hormone and substrate insulinogenic substances needs still to be elucidated. A fundamental question of urgent clinical dimension concerns the precise relationship of fetal hyperglycemia and hyperinsulinemia to perinatal morbidity and mortality. Empirical clinical obser- vations (34.,50,52) demonstrate that strict control of maternal hyperglycemia prevents the embry- opathy and improves the salvage rate of viable fetuses. Perinatal morbidity and mortality, however, even under these conditions, is still formidable. Even though considerable progress has been made in developing sophisticated laboratory aids (2,30,31,54,56), which help to signal fetus distress, the abnormalities uncovered still represent the trappings of a much more fundamental morbid process. The critical question of the mechanism of death in utero in these pregnancies is still to be elucidated. FETAL THERAPEUTICS The pathophysiologic sequence outlined above does provide a rationale for the management of pregnancy in the diabetic patient. The goal of fetal therapy is to prevent ketoacidosis with its attendent immediate and potential long-termed complications (12), and to minimize excessive pla- cental transfer of glucose to the fetus. In order to achieve this, maternal hyperglycemia is avoided by administering insulin in a manner calculated to simulate the demand response of the normal pancreas. To avoid postprandial surges in plasma glucose concentration, multiple dosages of intermediary and short-acting insulin alone or in combination through the day are required. The gradual evolution of effective treatment for the gravid diabetic has provided three im- portant lessons in fetal therapeutics. Simulation of the normal homeostatic adjustments of gestation may be the key to successful medical care. To reach this goal in the care of a woman with diabetes mellitus in pregnancy, the critical factor is careful composite chemical control (34). The criteria for optimum medical care in the gravid are different from those needed in the nongravid state. It is essential to understand the special exigencies and delicate balance of the diabetic who is pregnant, in order to practice optimal fetal therapeutics. An important derivative of this experience is the realization that total care of the mother and conceptus requires a team approach. Long and meticulous care of the maternal diabetic state, diagnostic amnicentesis, delivery and its timing, and definitive neonatal care require a large medical team composed of a knowledgeable referring physician, a diabetologist, obstetrician, neonatologist, trained biochemist, radiologist, nurses, social workers, and anesthesiologists. Survival is a function of the combined inputs of these professionals. The unfavorable gestational milieu accompanying diabetes mellitus represents a coersive influ- ence on judgements concerning optimum timing of delivery (66). Despite considerable investigational efforts in this direction, no laboratory aid has yet emerged which can serve as an immediate sensi- tive and specific prognosticator of fetal distress or fetal well-being. Many clinics have developed normograms of estriol levels (2,56,57), placental lactogen levels (30,56), and other direct or indirect indices of fetal function which are alleged to be effective in this critical assessment 260 Diabetes Mellitus process. Opinion, however, remains divergent as to their ultimate utility, and the need for continued investigation in this area of fetal diagnosis is still highlighted (19). The major cause of death in the newborn of a diabetic mother is the respiratory distress syndrome. It now appears that prediction and possibly prevention of this disease are imminent. The phospholipid lecithin is the major component of the surface active alveolar lining layer which determines alveolar stability in newborn life. The concentration of phospholipids in amniotic fluid is predictive of the potential for extrauterine alveolar stability. Gluck et al. (23, 24) correlated the lecithin sphingomyelin ratio in amniotic fluid with the subsequent inci- dence of respiratory distress syndrome. The relative lung maturity of the newborn may now be predicted by biochemical examination of a sample of amniotic fluid. In some circumstances, premature labor is imminent or must be induced because of overwhelming maternal or fetal risk even when the ratio of lecithin to sphingomyelin indicates a high proba- bility of functional immaturity in the newborn lung. DeLemos et al. (16) and Kotas et al. (37) in animal models, and subsequently Liggins et al. (40) in human premature infants, have presented evidence that fetal lung maturation can be therapeutically accelerated and prophylaxis thereby accomplished. In humans, glucocorticoids administered to the mother prior to 32-weeks gestation can increase the lecithin sphingomyelin ratio (58) and result in a significant reduction in the occurence of respiratory distress syndrome (40). This current experience opens the possibility for intrauterine fetal treatment prior to forced premature delivery in complicated diabetic and other high-risk pregnancies. Both short- and long-term side effects of this fetal therapy require prospective evaluation. Advances in fetal diagnosis and therapeutics are derivatives of the successful application of present knowledge. Continued advancement requires new information about the normal and pathological processes of pregnancy, the degree of control that the fetus-placenta exercises in regulating metabolic processes of the mother, the interdependence as well as independence of the fetal and maternal endocrine systems, and the time-related specificity of fetal metabolic pathways. Moreover, since fetal therapeutics, at its current state of development, is neither enzyme nor process specific, appraisal of the steps taken by the physician managing the fetus and its environment will require a focus more refined than the relatively gross standard of perinatal survival or mere absence of disease. CONCLUSION Considerable progress has been made in the last decade relative to our understanding of fac- tor(s) influencing fetal carbohydrate metabolism. Medical care which the pregnant patient with diabetes mellitus receives today should be a derivative of these physiological insights. The actual level of care, however, is frequently determined by factors other than the availability of relevant new physiological information. The patient's geography, the family socioeconomic status and level of education of the family unit, the psychodynamics of the pregnant patient with diabetes mellitus (39), and the proximity to centers devoted to special care are major limiting influences. In order to provide the optimal environment for the unborn child, the full scope of public health must be utilized. To do so at this time may require a reorientation of individual and national medical care and educational priorities. Effects on Fetal Growth and Development 261 { It was emphasized that ideal care for the individual pregnant, diabetic woman and her unborn child requires a consortium of health professionals. It is axiomatic that a physician responsible for only a few such patients in any single year will not be able to accumulate an essential clini- cal experience that matures his medical judgment. Centers of special care should be generally available to the physician for continuing education and to the patient for the expertise in clini- cal management that is often required. The predictable need of highly specialized neonatal care in some pregnancies makes mandatory predelivery referral to such centers if perinatal mortality/ morbidity is ever to be significantly reduced. A basic requirement for advancement in the emerging field of fetal therapeutics is new infor- mation about the normal and pathologic processes of pregnancy. The use of man as an experimental model is proscribed by ethical considerations, hence the search for animal models by investigators should receive continuing support. A parallel need of equal urgency is for support funds for training physicians as perinatal scientists. The last two decades have witnessed a significant decline in perinatal morbidity and mor- tality in pregnancies affected by maternal diabetes mellitus. Further improvement should result as increased effort is directed at both fetal and neonatal therapy. To accomplish this, however, will require expanded support for fundamental and applied research in fetal growth and development. REFERENCES 1. Adam, PA, K Teramo, N Raiha, D Gitlin, and R Schwartz 1969. Human fetal insulin metabolism early in gestation: Response to acute elevation of the fetal glucose concentration and placental transfer of human insulin I!3l, Diabetes 18:409. 2. Aleem, FA, JHM Pinkerton, and DN Neill 1969. Clinical significance of the amniotic fluid oestriol level. J Obstet Gynaecol Br Commonw 76:200. 3. Assan, R, and J Boillot 1971. Pancreatic glucagon and glucagon-like material in tissues and plasmas from human foetuses 6-26 weeks old. In Metabolic processes in the foetus and new born infant. Eds, JHP Jonix, HKA Visser, and JA Troelstra. Williams and Wilkins, Baltimore, p 193, 4. Beard, RW, RC Turner, and NW Oakley 1971. An investigation into the control of blood glucose in fetuses of normal and diabetic mothers. Proc Second Congress Perinatal Med, p 114, 5. Beck, P 1969. Progestin enhancement of the plasma insulin response to glucose in Rhesus monkey. Diabetes 18:146. 6. Beck, P, and WH Daughaday 1967. Human placental lactogen: Studies of its acute metabolic effects and disposition in normal man. J Clin Invest 46:103. 7. Bloom, SR, and DI Johnston 1972. Failure of glucagon release in infants of diabetic mothers. Brit Med J 4:453. 8. Chez, RA, and DH Mintz 1973a. The development and function of the human endocrine pancreas in the endocrine milieu of pregnancy, puerperuim, and childhood. Ross Laboratories, in press. 9. Chez, RA, DH Mintz, and EO Horger III 1970. Factors affecting the response to insulin in the subhuman pregnant primate. J Clin Invest 49:1517, 10. Chez, RA, DH Mintz, WA Reynolds, and DL Hutchinson 1973b. Maternal fetal plasma carbohydrate relationships in monkey pregnancy. In preparation. } 262 Diabetes Mellitus 11, 12, 13, 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24, 25. 26. 27. 28. 29. 30. Chinard, FP, V Danesino, WL Hartmann, ASG Huggett, W Paul, and SRM Reynolds 1956. The trans- mission of hexoses across the placenta in the human and the Rhesus monkey. J Physiol 132:289. Churchill, JA, HW Berendes, and J Newmoore 1970. Neuropsychological deficits in children of diabetic mothers. Pediatrics 105:257. Clark, MC, Jr, B Beatty, and DO Allan 1973. Evidence for delayed development of the glucagon receptor of adenylate cyclase in the fetal and neonatal rat heart. J Clin Invest 52:1018. Cornblath, M, and R Schwartz 1966. Disorders of carbohydrate metabolism in infancy. WB Saunders Company, Philadelphia, p 82. DeGaspero, M, and JJ Hoet 1970. Normal and abnormal foetal weight gain. Proc of the VII Congress International Diab Fed. In Exerpta Medica International Congress Series No 231, Buenos Aires. deLemos, R, DW Shermeta, J Knelson, RV Kotas, and ME Avery 1970. Acceleration of appearance of pulmonary surfactant in fetal lamb by administration of corticosteroids. Amer Rev Resp Dis 102:459. Falorni, A, F Fracassini, F Massi Benedietti, and A Amici 1972. Glucose metabolism, plasma insulin, and growth hormone secretion in newborn infants with erythroblastosis fetalis com- pared with normal newborns and those born to diabetic women. Pediatrics 49:682. Farquhar, JW 1971a. Diabetes and the fetus: Questions which need answers. CMA Jornal 105:289. Farquhar, JW 1971b. The antenatal treatment of the fetus of the diabetic mother. CMA Journal 105:287. Fee, BA, and WB Weil 1963. Body composition of infants of diabetic mothers by direct analysis. Ann NY Acad Sci 110:869. Freinkel, N 1965. Effects of the conceptus on maternal metabolism during pregnancy. In On the nature and treatment of diabetes. Eds, BS Liebel and J Wrenshall. Excerpta Medica International Congress Series No 84, Moutan and Company, Netherlands. Freinkel, N, E Herrera, RH Knopp, and HJ Ruder 1968. Metabolic realignment in late pregnancy. Adv in Met Dis, Suppl 1, 205. Gluck, L, and MV Kulovich 1973. Lecithin/sphingomyelin ratios in amniotic fluid in normal and abnormal pregnancy. Amer J Obstet Gynecol 115:539. Gluck, L, MV Kulovich, RC Borer, PH Brenner, GG Anderson, and WN Spellacy 1971. Diagnosis of the respiratory distress syndrome by amniocentesis. Amer J Obstet Gynecol 109:443. Grasso, S, A Messina, N Saporito, and G Reitano 1970. Effect of theophylline plus glucagon in insulin secretion in the premature infant. Diabetes 19:837. Grasso, S, N Saporito, A Messina, and G Reitano 1968. Serum-insulin response to glucose and amino acids in the premature infant. Lancet 2:755. Green, JW, Jr, and JL Duhring 1971. Diabetes and pregnancy. J Tenn Med Asso 64:113. Hommes, FA, and A Beere 1971. The development of adenyl cyclase in rat liver, kidney, brain and skeletal muscle. Biochim Biophys Acta 237:296. Johnston, DI, SR Bloom, KR Greene, and RW Beard 1972. Plasma pancreatic glucagon relationship between mother and fetus at term. J Endo 55:25. Josimovich, JB, B Kosor, L Bocella, DH Mintz, and DL Hutchinson 1970. Placental lactogen in maternal serum as an index of fetal health. Obstet and Gynecol 36:244. 31. 32. 33. 34. 35. 36. 37. 38. 29. 40. 41. 42. 43. 44, 45. 46. 47. 48. 49. 50. Effects on Fetal Growth and Development 263 Kaiser, E, and H Crabben 1969. Neuer hormon-belastungstet bei plazenta-insuffizienz. Seventh International Kongresz fuer Klinische Chemie, Geneva. Kalkhoff, RK, BL Richardson, and P Bech 1969. Relative effects of pregnancy, human placental lactogen and prednisolone and carbohydrate tolerance in normal and subclinical diabetic subjects. Diabetes 18:153. Kaplan, SL, MM Grumbach, and TH Shepard 1972. The ontogenesis of human fetal hormones. I. Growth hormone and insulin. J Clin Invest 51:3080. Karlsson, K and I Kjillmer 1972. The outcome of diabetic pregnancies in relation to the mother's blood sugar level. Amer J Obstet Gynecol 112:213. King, KC, PA Adam, GA Clemente, and R Schwartz 1969. Infants of diabetic mothers: Attenuated glucose uptake without hyperinsulinemia during continuous glucose infusion. Pediatrics 44:381. King, KC, J Butt, K Raivo, N Raiha, J Roux, K Teramo, K Yamaguchi, and R Schwartz 1971. Human maternal and fetal insulin response to arginine. New Engl J Med 285:607. Kotas, RV, and ME Avery 1971. Accelerated appearance of pulmonary surfactant in the fetal rabbit. J Appl Physiol 30:358. Kyle, GC 1963. Diabetes and pregnancy. Ann Int Med 59 (suppl):1. Leeman, CP 1970. Dependency, anger, and denial in pregnant diabetic women: A group approach. Psych Quart 44:1. Liggins, GC, and RN Howie 1972. A controlled trial of antepartum glucocorticoid treatment for prevention of the respiratory distress syndrome in premature infants. Pediatrics 50:515. Luyckx, AS, F Massi-Benedetti, A Falorni, and PJ Lefebrore 1972. Presence of pancreatic glucagon in the portal plasma of human neonates. Difference in the insulin and glucagon response to glucose between normal infants and infants from diabetic mothers. Diabetologia 8:296. Metzger, B, JW Har, and N Freinkel 1972. A new role for the placenta: Nitrogen conservation during pregnancy. Diabetes 21:340. Metzger, B, JB Paton, DE Fisher, and N Freinkel 1973. Selective increase in pancreatic blood flow during pregnancy and beta cell function. Diabetes 22:296. Milner, RDG 1971. The development of insulin secretion in man in metabolic processes in the foetus and newborn infant. Eds, JHP Jonix, HKA Visser, and JA Troelstra. Williams and Wilkins, Baltimore, p 193. Milner, RDG, and CN Hales 1965. Effect of intravenous glucose in concentration of insulin in maternal and umbilical-cord plasma. Brit Med J 1:284. Mintz, DH, RA Chez, and EO Horger III 1969. Fetal insulin and growth hormone metabolism in the subhuman primate. J Clin Invest 48:176. Mintz, DH, RA Chez, and DL Hutchinson 1972. Subhuman primate pregnancy complicated by streptozotocin-induced diabetes mellitus. J Clin Invest 51:837. Mintz, DH, R Stock, JL Finster, and AL Taylov 1968. The effect of normal and diabetic preg- nancy on growth hormone responses to hypoglycemia. Metabolism 17:54. Mglsted-Pedersen, L, and KR Jdrgensen 1972. Aspects of carbohydrate metabolism in newborn infants of diabetic mothers III: Plasma insulin during intravenous glucose tolerance test. Acta Endo 71:115. Oakley, W 1965. The treatment of pregnancy in diabetes mellitus. In On the nature and treatment of diabetes. Eds, BS Leibel and GA Wrenshall. Excerpta Medica International Congress Series No 84, Moutan and Company, Netherlands 264 51. 52. 53. 54. 55, 56. 57. 58. 59. 60. 61. 62. Diabetes Mellitus O'Sullivan, JB, and CM Mahan 1964. Criteria for oral glucose tolerance test in pregnancy. Diabetes 13:278. Pedersen, J 1954. Foetal mortality in relation to management during the latter part of pregnancy. Acta Endocr (Kbh) 15:282. Pedersen, J 1967. Pregnant diabetic and her newborn: Problems and management. Williams and * Wilkins Co, Baltimore Rivlin, ME, JH Mestman, TD Hall, CP Weaver, and GV Anderson 1970. Value of estriol estima- tions in management of diabetic pregnancy. Amer J Obstet Gynecol 106:875. Robb, P 1961. The development of the islets of Langerhans in the human fetus. Quart J Exp Physiol 46:335. Samaan, NA, HS Gallager, WA McRoberts, and AM Faris, Jr 1971. Serial estimation of human placental lactogen, estriol, and prenanediol in pregnancy correlated with whole organ section of placenta. Amer J Obstet Gynecol 109:63. Schwarz, RH, and GA Fields 1971. The management of the pregnant diabetic. Obstet Gynecol Survey 26:277. Spellacy, WN, WC Buhl, FC Riggalls, and KL Holsinger 1973. Human amniotic fluid lecithin/ sphingomyelin ratio changes with estrogen or glucocorticoids treatment. Amer J Obstet Gynecol 115:216. Thomas, K, M DeGaspero, and JJ Hoet 1967. Insulin levels in the umbilical vein and in the umbilical artery of newborns of normal and gestational diabetic mothers. Diabetologia 3:299. Van Assche, FA, W Gepts, M deGaspero, and M Renaer 1970. The endocrine pancreas in erythro- blastosis fetalis. Biol Neonat (Basel) 15:176. White, P 1971. Pregnancy and diabetes. In Joslin diabetes mellitus. Eds, A Marble, P White, R Bradley, and L Krall. Lee Febiger, Philadelphia. Williger, VM 1966. Fetal outcome in diabetic pregnancy. Amer J Obstet Gynecol 94:57. Chapter Chapter Chapter Chapter Chapter Chapter 20 21 22 23 24 25 NEW APPROACHES TO CONTROL AND PREVENTION Diabetes Mellitus: A Bioengineering Approach-- An Implantable Glucose Sensor J. S. Soeldner, K. W. Chang, Sol Aisenberg, J. M., Hiebert, and R. H. Egdahl Transplantation of Insulin-Secreting Tissues R. C. Karl, D. W. Scharp, Paul E. Lacy, and W. F. Ballinger Diet and Diabetes Mellitus Ronald K. Kalkoff Insulin Synthesis and Analogs Harold E. Lebovitz Drugs Enhancing Insulin Secretion Harold E. Lebovitz Drug Altering Carbohydrate and Lipid Metabolism Bernard Robert Landau 267 278 295 310 327 344 20 DIABETES MELLITUS : A BIOENGINEERING APPROACH--AN IMPLANTABLE GLUCOSE SENSOR J. Stuart Soeldner, Kuo Wei Chang, Sol Aisenberg, John M. Hiebert, and Richard H. Egdahl ABSTRACT There is increasing concern focused upon diabetes mellitus and its complications. Current evidence strongly suggests that a desired endpoint is the normalization of blood glucose levels. Available therapeutic programs rarely allow the diabetic to achieve normal glucose levels due to the inability of a program to either frequently monitor blood glucose levels or to mimic the normal dynamic function of the beta cells of the Islets of Langerhans in the pancreas. A project has been initiated which focuses on the construction of a small implant- able glucose electrode which can be incorporated into an implantable "Glucose Monitor' and which could provide the diabetic patient with frequent, instantaneous information on blood glucose control. It would produce an audible and/or visual alarm if blood glucose rose above or below a desired pre-set level. After further development, the glucose electrode could be incorporated into an implanted "Artificial Beta Cell" which would be a demand insulin release system pro- grammed for normalization of blood sugar. INTRODUCTION Diabetes Mellitus is a disease which presents many problems and many challenges. Although it is the fifth leading cause of death, it is thought by many that the death rate from diabetes is underestimated. The major reason for this is that the complications of diabetes, particu- larly those that affect the cardiovascular system are often considered as the primary cause of death while diabetes itself is ignored or listed as a minor cause. Another important considera- tion concerning diabetes is the fact that at least 4 percent of the population of the United States have the disease in an overt or occult form. The major complication of diabetes is coronary heart disease which in diabetics causes 3 to 4 times as many deaths than in the nondiabetic general population. In addition, it is recognized that death due to heart complications in diabetics occurs at a younger age than in the nondiabetic. Other complications include renal failure which is nearly 20 times more common in the diabetic patient than in the nondiabetic. The second leading cause of blindness is diabetes. Numerous other complications including gangrene, neuropathy, infections, etc., also serve to reduce the quality and duration of life in the diabetic. CURRENT CONCEPTS OF THE COMPLICATIONS Much evidence suggests that the frequency of the complications is reduced in patients with good control of blood sugar (glucose) (13). More current evidence supports the concept that Supported by the Doll Foundation, New York City, N.Y., and the Joslin Diabetes Foundation, Inc., Boston, Mass. 267 268 Diabetes Mellitus elevated blood glucose levels may by itself induce metabolic and structural changes (particularly in small and large blood vessels) which underlie the broad spectrum of diabetic complications. Although the precise mechanisms have not yet been clarified, greater focus has been placed upon systems and techniques designed to truly normalize blood glucose levels in the diabetic patient. CURRENT DIABETIC MANAGEMENT The keystones of a diabetic treatment program today are essentially diet, exercise, and specific antidiabetic therapy (i.e., insulin injections for the more severe and diet alone or combined with oral antidiabetic agents in the milder forms of the disease). Currently available control monitoring systems are crude considering the complexity and the dynamics of blood glucose homeostasis. Sporadic blood glucose determinations (perhaps only once a month) provide the diabetic patient and his physician with only a fleeting retrospective view of the degree of control. Urine glucose determinations are simpler than blood glucose deter- minations, can easily be employed by the patient, and can be as frequent as urination. However, certain features of the physiology of urine excretion of glucose make this determination gross at best. The main problems focus on the fact that there is a "threshold" of blood glucose con- centration below which no glucose can be detected in the urine. Usually this threshold is 160 mg per 100 ml of blood glucose, but this tends to vary from person to person and increases with age. In addition, this determination of urine glucose is an integrated value, only crudely correlating with blood glucose and not indicating whether the elevated blood glucose is rising, stable, or falling. Finally, the urine test is negative for glucose at blood glucose values below threshold. These apparently negative values could be higher or lower than the range of blood glucose seen in healthy nondiabetic subjects (60 to 140 mg per 100 ml). A GLUCOSE ELECTRODE It appears conceivable that a small, inert, nontoxic, biocompatible, glucose sensing device which requires low power would have many practical applications in relation to diabetes control and could be the critical component of a mechanical device focused upon improving blood sugar control in diabetics. Updike and Hicks (16) and Hicks and Updike (12) described a '"glucose sensor' constructed around a Clark (7) oxygen probe. They introduced glucose oxidase enzyme in a film of hydrogel (polyacrylamide) positioned directly over the teflon membrane covering the platinum cathode. Glucose added to solutions in which the device was inserted was oxidized and the residual PO, activating the oxygen electrode reduced. The current output of the electrode fell in a nonlinear fashion proportional to the glucose concentration. A similar type of electrode had been proposed by Clark and Lyons (8) and Clark and Sachs (9). Irreversible loss of enzyme activity in these systems rendered them unsuitable for long-term use. Bessman and Schultz (2) extended these studies. In their version, the glucose oxidase enzyme is covalently bound to a rayon acetate cloth and a special spiral type of silver cathode, lead anode type oxygen electrode was developed. Two electrodes are used, both electrode faces are covered by a membrane, onto one of which the enzyme oxidase has been conjugated. This arrangement allows for the oxidation of glucose in the area of one of the oxygen electrodes and consequently a lowering of the PO, relative to the other oxygen electrode. The difference in PO, between the electrodes is related to the glucose concentration. A Bioengineering Approach 269 Additional studies by Bessman and Schultz (1) had focused upon a regeneratable noble metal catalyst as a substitute for glucose oxidase. These studies demonstrated catalyst poisoning that was partially circumvented by an electric pulsing system. A survey by the Joslin Research Laboratory group of methods and techniques used for measure- ment of glucose indicated that a new and novel method would have to be devised for long-term implantable use (4). An intriguing possibility was found in the area of glucose electrochemistry since glucose oxidation can be catalyzed in the presence of certain noble metals (3,14). The general scheme of postulated reactions that take place at the cathode and anode of a glucose sensor is: + Anode: 2CgH;20g + 2H20 —s Pt* ———s2CgH;207 + 4H + 4e Cathode: 0, + 4H + 4e — Pt* ——= 40H" Overall: 0, + 2CgH; 206 — Pt» — 2CgH; 207 *Platinum electrode catalyst A program for the construction of a glucose electrode evolved which led to the fabrication of - the device shown in Fig. 1. STAINLESS STEEL CLAMPING PLATE OXYGEN PERMEABLE MEMBRANE (Sincone Rubber 0.0005") OXYGEN CATHODE (Plotinum Block 0.003") FUEL ANODE (Platinum Black 0.003") GASKET (Silicone Rubber 0.027") GASKET (Silicone Rubber 0.027") ION EXCHANGE MEMBRANE (AMF lon Exchonge Membrane 0.007") SEMI-PERMEABLE MEMBRANE (Cellulose 0.0008") PULSING ELECTRODE (Tontalum Screen 0.006") STAINLESS STEEL CLAMPING PLATE FIGURE 1. A schematic of the glucose sensor. The membrane covering the fuel anode compartment was specially selected and constructed not only to shield the electrode from high molecular weight species, but to allow diffusion of glucose, oxygen, and water. The membrane over the oxygen cathode was selected on the basis of allowing 02 and H,0 vapor diffusion but excluding glucose and other nonvolatile, water-soluble species. The ion exchange membrane placed between the two electrodes serves as a solid-state ion shuttle. 270 Diabetes Mellitus ELECTRODE PERFORMANCE Prototype and modified electrodes have been studied extensively in vitro and in vivo (5,15, 6). The in vitro studies demonstrated that the device was stable and showed an increase in cur- rent as a function of glucose concentration (Figs. 2 and 3). 15 T T le— BATH ADJUSTMENT PERIOD ————, ee... 10+ 7 . ot) sesasstecssssnnsenes ea ate s0eteests yy oeeenes,. S I FUNGUS AND BACTERIA 0.05% NgN3, 0.05% N,F, 0.02% LINCOCIN ,0.001%PANMYCIN + INHIBITORS: ceestesrategtest eSROetesy tuete Case t tte ts tnsatte, |, eansieenets teen, tees eresenssanecaanie esarane . eaten, s000 o SENSOR CURRENT (uA) eosentssonct o oe sang, 0000rse0e® a %sa000 te 20 00000 e000s tue ete #400" %0%e0e% tanruasten met $9000 etuarsananntasassnenas® ts 0 APPEARANCE OF FUNGUS ¢ BACTERIA 1 5 6 7 TIME ELASPED (days) FIGURE 2. Results of a 7-day study in which a glucose sensor was incu- bated in a Krebs-Ringer buffer containing glucose (100 mg per dl) and a mixture of antibiotics. Sensor readings were recorded every 15 minutes. The overall coefficient of correlation was 6.6 percent (5). o In addition, it had a high specificity for glucose compared to other compounds found in biologi- aal fluids and could be constructed to respond rapidly to abrupt changes of glucose concentration. Preliminary in vivo studies were performed in a variety of animals (monkeys, rabbits, and dogs). In these studies, the glucose electrode was attached via wires to an external power supply, amplifier, and recording unit (Fig. 4). The glucose electrode was implanted into subcutaneous tissues (abdomen or back) and in studies to date has survived for up to 117 days and produced a signal which correlates significantly with corresponding blood sugar levels following intravenous glucose administration (Fig. 5). Tn the majority of these trials, fracture of wires and infection associated with the skin area in which the wires emerged was the cause of the termina- tion of the study. Current efforts are being focused upon the development of a totally implantable unit containing the glucose electrode, power supply, and a miniature radio transmitter with a suit- able external receiver-recorder system (Fig. 6). This modification will be employed for more prolonged (6 months plus) studies of the performance characteristics of the electrodes. Following the development of a satisfactory totally implantable system and documentation of the accuracy, reliability, and reproducibility of the glucose electrode signal in animals, trials in diabetic patients will be started. A Bioengineering Approach 271 1000 T T rrr" T T T TTY T TOTAL NUMBER OF MEASURMENTS : 411 1 MEASUREMENT PERIOD: FROM [Oth TO 15th DAY ] BACTERIA § FUNGUS INHIBITORS: 0.05% NgN3 0.05% NgF ! 0.02% LINCOCIN J 0.001% PANMYCIN 2, 100 — CELL CURRENT DENSITY (pA /CM NORMAL RANGE IN HUMANS 10 1 Ld eta a 1 1 L aa) 20 50 100 200 500 1000 2000 DEXTROSE CONCENTRATION (mg/100mi) FIGURE 3. The dose response pattern of a glucose sensor incubated in Krebs-Ringer buffer at glucose concentrations ranging from 50 to 1000 mg per dl. Amplifier— Chart Recorder Glucose Electrode FIGURE 4. This shows the arrangement employed for the initial animal studies. 272 Diabetes Mellitus 60 o 200}— h ] MONKEY GS6 - 4.6 Kg $ —- ! § C no DAY 117 SENSOR No. 103 — ~ \ Poi-2 | | z £E I ; ~ I z ial ! SENSOR CURRENT Eso 8 [ : 3 2 fF | BLOOD GLUCOSE © — I «x °o ! B Bat ] W sol. 39° re ! i» ! i 4 20 50 [ 1 IVGTT 1 Loni 1 le J 14:00 15:00 16:00 17:00 TIME (Hours) FIGURE 5. This shows a comparison of the glucose sensor current and the blood glucose levels obtained simultaneously during an intra- venous glucose tolerance test in a female Rhesus monkey after 117 days of implantation (6). APPROACHES TO BETTER DIABETES CONTROL A general plan has been formulated by this group and a series of stages have been developed. Stage I - A Glucose Monitor It is plausible that the great majority of diabetic patients could be educated and trained to alter diet, exercise, and antidiabetic medication to produce better blood glucose control Zf they had constantly available information as to their body's glucose level. We propose initially to construct a device suitable for implantation into body tissue which could sense the glucose level and transmit this signal externally via an appropriate telemetry system to a suitable ex- ternal receiver-recording device (Fig. 6). Experience to date has been confined to implantation in extravascular sites and the extent to which they reflect instantaneously changes in blood sugar remains to be clarified. This reading would be available to the diabetic patient on a moment-to-moment basis and could guide in the proper selection or modification of his regime. Also, the external receiver device could be constructed to produce an appropriate audible and/or visible signal if glucose levels rose above or fell below preset limits. The latter mode of operation would be particularly valuable for the unstable juvenile type of diabetic who is prone to develop serious episodes of insulin reaction (low blood glucose) at night while sleeping. Thus alerted, the diabetic could consume an appropriate amount of carbohydrate and avert the reaction. Stage II - The Artificial Implantable Beta Cell This stage will be difficult, but the validation of the implanted glucose electrode required for Stage I will obviously solve one of the problems. A second component of the artificial beta cell will be a miniature computer which can translate the signal from the electrode to an appropriate amount of insulin release (Fig. 7). A Bioengineering Approach 273 Glucose Electrode ” i eed) / - oe _. 7 oi 1 - . / -N.-{.Receiver- SN]. Alarm Power Supply Transmitter FIGURE 6. plantable "Glucose Monitor' (15). from Plenum Press. This depicts the major components of a totally im- Reprinted with permission So a! Glucose Electrode Supply Power Pump Insulin Computer Reservoir FIGURE 7. This depicts the major compo- nents of a totally implantable "Artificial Beta-Cell" (15). Reprinted with permission from Plenum Press. upon the temporal dynamics of glucose stimulated insulin release. It Preliminary studies have been done in modeling glucose homeostasis in man with particular emphasis appears that the Systems Dynamics Model of R. 0. Foster (10) as well as the model proposed by Grodsky (11) can be imple- mented in terms of a small computer component of the artificial beta cell. 274 Diabetes Mellitus An insulin reservoir will be an important component. Currently, this is conceived as a solid tank except for one surface which would be constructed of a self-sealing membrane. During implantation, the reservoir would be positioned such that the membrane would face the skin. To refill, the patient would fill a syringe with insulin; then after skin cleansing, he would pass the needle through the skin, subcutaneous tissue, and the self-sealing membrane until the tip was positioned inside the tank prior to injecting the insulin. Power requirements for both the Stage I Glucose Monitor and the Stage II Artificial Beta Cell could be in the form of batteries currently being used in cardiac pacemakers. Recently, atomic power supplies have been introduced into pacemakers as have externally rechargeable batteries. It is also conceivable that a ''glucose fuel cell" could be utilized (17). Here, design would allow for a constant current at all glucose levels to provide power for the unit (Fig. 8). GENERAL DISCUSSION It is difficult to chart a time table for both Stage I and II. Few precedents in biomedical devices are available to act as guidelines. Currently (June 1973) prototype "Glucose Monitors" are being constructed for prolonged evaluation totally implanted in animals. Figure 9 shows the critical path toward the Glucose Monitor. Success in this phase will be achieved when at least 18 or 20 of the devices perform satisfactorily for 6 or more months. At that point, it will be- come feasible to undertake clinical trials to evaluate whether normalization of blood sugar can materially reduce the incidence of diabetic complications. The timetable for the "Artificial Beta Cell" is obviously more complex. Figure 10 shows the critical path toward an Artificial Implantable Beta Cell. It is anticipated that the valida- tion of the performance of the glucose electrode will have been completed, and the major focus during this stage will be placed upon the insulin reservoir, pump, and computer interface. This will involve considerable skills in design, engineering, and physiology. . Fl A | Receiver- coo UN]. Alarm Glucose Electrode Transmitter £io1"Cel| FIGURE 8. This depicts the manner in which a glucose fuel cell could provide power for an implanted '"Glucose Monitor." A Bioengineering Approach 275 SENSORS IN VITRO AND INTERNAL ELECTRONICS ANIMAL TESTS EXTERNAL ELECTRONICS IMPLANTABLE SENSOR-MONITOR . IMPLANTABLE IMPLANT SENSOR- MONITOR TESTS SYSTEM FOR (Humans) ° HUMANS IMPLANT ALARM SYSTEM BIO-MATERIALS TESTS (Animals) PROGRAM FLOW DIAGRAM For Implantable Sensor-Monitor System for Humans FIGURE 9. The Program Flow Diagram for the implantable "Glucose Monitor' for humans. { 1 IMPLANTABLE SENSOR - MONITOR IN VITRO AND INPLaNT SYSTEM FOR ANIMAL TESTS HUMANS (Humans) INTERNAL ELECTRONICS IMPLANTABLE IMPLANT SENSOR-MONITOR a4 TESTS ALARM SYSTEM (Animals) EXTERNAL P=] ELecTRONICS INTERNAL INTERNAL IMPLANT INTERNAL IMPLANT Ek DISPENSER = A ceLL |— TESTS B CELL [—= TESTS ¢£ ELECTRONICS (Animals) (Animals) (Humans) (Humans) BIO-MATERIALS IMPLANTABLE B CELLS FOR n CLINICAL HUMANS i SLucost MONITOR SIMPLIFIED ELECTRONICS T AUTOMATIC % og |] ANIMAL |] EX TENA HUMAN EXTERNAL (Animals ) TESTS (Humans) TESTS 8 CELL FOR CLINICAL USE SYSTEM CONTROLLER PROGRAM FLOW DIAGRAM EXTERNAL , hn ERNAL OR For Extension to Artifical 8 Cell. FIGURE 10. The Program Flow Diagr am for the "Artificial Beta-Cell" in humans. 276 Diabetes Mellitus FINAL PRACTICAL CONSIDERATIONS The greatest question that still remains to be answered is: Does better control or indeed normalization of blood glucose dynamics in the diabetic reduce or eliminate the so-called diabetic complications? It is evident that solid scientific proof is still lacking. It also appears that until therapeutic systems are available for treating diabetes that are much superior to any cur- rent method, then a sound broad-based prospective study could not be performed. A second question that arises (if lack of blood glucose control does relate to the complica- tions) would address the problem of establishing some criteria whereby if a diabetic's degree of control was not '"satisfactory'" with conventional measures, then the employment of a device (i.e., Glucose Monitor or Artificial Beta Cell) would be warranted. It is apparent that there are certain diabetics who would qualify even with the current uncertainty concerning the long-term consequences of control. These types of diabetics would include those with frequent episodes of ketoacidosis, frequent hypoglycemic reactions especially without warning symptoms, the brittle-unstable diabetic, and probably those with a high renal threshold. REFERENCES 1. Bessman, SP, and RD Schultz 1972. Sugar electrode for the "Artificial Pancreas." Horm Metab Res 4:413. 2. Bessman, SP, and RD Schultz 1973. Prototype glucose-oxygen sensor for the artificial pancreas. Trans Amer Soc Artif Int Org 19:361. 3. Bockris, JO'M, and S Srinivasan 1969. Fuel Cells: Their Electrochemistry, McGraw-Hill, New York. 4. Cahill, GF Jr, JS Soeldner, GW Harris, and RD Foster 1972. Practical development in diabetes research. Diabetes 21:703. 5. Chang, KW, S Aisenberg, and JS Soeldner 1972. In-vitro tests of an implantable glucose sensor. Proc of 25th Ann Conf on Eng in Med and Biol, p 58. 6. Chang, KS, S Aisenberg, JS Soeldner, and JM Hiebert 1973. Validation and bioengineering aspects of an implantable glucose sensor. Trans Amer Soc Art Internal Org 19:352. 7. Clark, LC, Jr, 1956. Monitor and control of blood and tissue oxygen tensions. Trans Soc Art Int Org 2:41. 8. Clark, LC, Jr and C Lyons 1962. Electrode systems for continuous monitoring in cardio- vascular surgery. Ann NY Acad Sci 102:29. 9. Clark, LC, Jr and G Sachs 1968. Bioelectrodes for tissue metabolism. Ann NY Acad Sci 148: 133. 10. Foster, RO 1970. The dynamics of blood sugar regulation. Thesis, Massachusetts Inst of Tech. 11. Grodsky, GM 1972. A threshold distribution hypothesis for packet storage of insulin and its mathematical modeling. J Clin Invest 51:2047. 12. Hicks, GP and SJ Updike 1970. Enzyme electrode. US Patent 3,542,662:1. 13. Joslin's Diabetes Mellitus 1971. Eds, A Marble, P White, LP Krall, and RF Bradley. Phila- delphia, Lea and Febiger. 14. 1S. 16. 17. A Bioengineering Approach 277 Mitchell, W Jr 1963. Fuel Cells, Academic Press, New York. Soeldner, JS, KW Chang, S Aisenberg, and JM Hiebert 1973. Progress toward an implantable glucose sensor and an artificial beta cell. In: Temporal Aspects of Therapeutics. Proceed- ings of the Second Alza Research Conference, Eds, J Urquhart and FE Yates. Plenum Press, New York-London, pp 181-207. Updike, SJ and GP Hicks 1967. The enzyme electrode. Nature 214:986. Wolfson, SK, Jr, and CK Strohl Jr 1967. Bioautofuel cell as a possible power source for cardiac pacemakers. Circulation 36 (Suppl 2):273. 21 TRANSPLANTATION OF INSULIN SECRETING TISSUES Richard C. Karl, David W. Scharp, Paul E. Lacy, and Walter F. Ballinger There are between 3 and 5 million individuals in the United States suffering from diabetes mellitus, and it is estimated that another 5 million, now alive, will develop the disease during their lifetime. Diabetes is probably the most common metabolic disorder in the western world today. The discovery of insulin in 1921 and its extraction and preparation for clinical use shortly thereafter was one of the most dramatic therapeutic breakthroughs achieved by modern bio- logical science. It has been estimated that in 1916, 64 percent of the deaths of diabetic pa- tients were due to diabetic coma. At that time, a child diagnosed as a juvenile diabetic was ex- pected to be dead within two years. Today diabetic coma accounts for less than 1 percent of diabetic deaths and juvenile diabetics are still setting longevity records. Yet, the increased longevity of both adult and juvenile onset diabetics made possible by insulin therapy and the advent of antimicrobials has revealed a host of complications thought to be secondary to the metabolic disorder. Both macro and microvascular disease is accelerated in the diabetic and these processes affect the brain, heart, kidneys, eyes, and the extremeties. The dismaying impotence of accepted therapeutic approaches to alter the increased morbidity and mortality caused by this secondary vascular disease has recently been quantitated by the University Group Diabetes Program (2). Although there are some legitimate objections to this study (3), it is universally accepted that insulin and the oral hypoglycemic agents have not rendered the diabetic patient normal. Other chapters of this monograph discuss in greater detail the complica- tions of diabetes and the currently employed therapeutic agents. The question of whether the vascular changes seen in diabetes are secondary to the meta- bolic derangement, or are instead manifestations of an entirely independent disorder which happens to be co-inherited with diabetes, is not yet fully resolved. Nonetheless, persuasive evidence has been put forward that indicates that the complications seen in diabetics are consequences of the disease. Williamson has quantitated changes in capillary basement membrane thickness in normals and diabetics (4,5,6). He has found that in both normals and diabetics this sign of microvascular disease is increased with age. Observations in diabetics indicate that these patients probably had normal capillary basement membrane thickness prior to the onset of their disease. Patients with the complications of retinopathy or nephropathy have a 90 percent incidence to capillary base- ment membrane thickening. Patients suffering from the disease for 20 years or longer have a 93 percent incidence of such thickening. These data suggest that the metabolic derangement precedes the abnormality of the capillary basement membrane. Furthermore, the occurrence of retinopathy and nephropathy in patients suffering from nonhereditary forms of diabetes such as hemochromatosis (7) or pancreatic damage (8) is suggestive of a causal relationship between diabetes and vascular disease. Although great controversy has been elicited about the exact biochemical lesion in diabetes, it is reasonably well accepted that at least one basic defect resides in the B cell of the islets of Langerhans (9). Thus, all forms of diabetes are characterized by a "relative" deficiency of 278 Transplantation of Insulin Secreting Tissues 279 circulating insulin (10). ‘Detailed studies of insulin secretion in normal and diabetic subjects indicate that the cells ar¢: sluggish in their response to glucose stimulation (9). The failure of exogen.ous agents (e.g., insulin and the oral hypoglycemics) to effect a rigorous control of blood sugar from minute to minute may explain the failure of such agents to prevent the complications; of diabetes. As the basic defect in diabetes lies in the B cells, it is postulated that successful transplantation of normal B cells into the diabetic patient may prevent the development of these crippling complications. MODELS FOR TRANSPLANTAT ION Experiments desigiaed to test the feasibility of transplanting functioning insulin secreting tissue capable of "curring" diabetes all have a common requirement: the recipient animal must be diabetic. Although a. few experiments have been done involving spontaneously diabetic animals, almost all investiga tions have employed animals rendered diabetic by the administration of a specific cell poisori, alloxan, or streptozotocin. The biochemical mode of action of both these agents is obscure. Alloxan is thought to be the less reliable of the two, as it is labile at room temperature and late reversion of alloxan diabetic animals to normal has been reported. Nonetheless both these agents have been very important tools in the evaluation of the efficacy of B cell transp lantation. TRANSPLANTATION OF THE PANCREAS The role /of pancreatic islets in the pathogenesis of diabetes was under active investigation by the late nineteenth century. In 1889 Von Mering and Minkowski first demonstrated that hyper- glycemia foll owed pancreatectomy (11). In 1892 Hendon reported a technique of placing a vascu- larized auto graft of a segment of the pancreas in the subcutaneous tissues of dogs (12). When the remainirig nontransplanted pancreas tissue was excised, the dogs did not develop hyperglycemia. In light of such observations Ssobolew first proposed pancreatic transplantation as a cure for diabetes i'n 1902 (13). Atternipts at pancreatic transplantation for the cure of diabetes have involved the use of the whole gla nd or portions of the pancreas, with or without direct vascular anastomosis between the graft and the recipient's vascular system. In addition, pancreas fragments and isolated islets of Lang¢:rhans have been transplanted. TRANSPI _ANTATION OF THE PANCREAS WITH VASCULAR ANASTO MOSES IN EXPERIMENTAL ANIMALS "Transplantation of the entire pancreas or portions of the pancreas with vascular anastomoses to tlae recipient arterial and venous system have been described for many years. The recipient anifnals were usually rendered diabetic by pancreatectomy or the administration of alloxan. Most investigators described only a modest survival rate in homotransplant recipients. Some of those animals surviving transplant have demonstrated endocrine function of the graft for varying short p2:riods of time. In 1966 Largiader reported transplantation of the entire pancreas into un- related dogs (14). The pancreas and duodenum were transplanted enbloc and the duodenum was Clrained into the gallbladder. Approximatley 50 percent of the animals succumbed immediately post- operatively; the remainder survived only four to nine days. These short survivors maintained normoglycemia until death. Unfortunately, no circulating insulin levels were measured. In a series reported by Merkel, one-half of surviving canine whole pancreas recipients had evidence of 280 Diabetes Mellitus functioning endocrine tissue. This was indicated by near normal serum g'lucose levels and elevated circulating insulin levels when compared to diabetic controls (15). In :inother set of experi- ments, 33 percent of those diabetic dogs receiving pancreatic allografts were normoglycemic for a period up to 32 days (16). Some of these dogs demonstrated significant increases in insulin secretion rate when challenged with a glucose tolerance test. These homotransplantation studies in dogs unfortunately combine problems of tissue rejection with the technical challenges posed by such surgery. A study of Aquino et &il., using an auto transplant model (thereby eliminating tissue incompatibility phenomena), demonstrated that the surgical problems caused a significant percentage of failures (17). Only 2 of 12 dogs survived longer than 17 days. Mortality of the remaining dogs was due to vascular thrombosis, anastomotic leaks, and intercurrent infection (see below). Studies in inbred strains of rats have also been designed to circumvent tissue incompatibility complications. The microsurgery necessitated by the caliber of rat blood vessels is difficult, and a 48-hour postoperative mortality rate of 60 percent has been described for pancreas transplantation (18). A few inbred rat pancreas recipients did survive and were normoglycemic. Resection of the transplanted graft was followed by hyperglycemia. In a recent study Lee et al. described isologous pancreato-duodenal transplants in inbred Lewis rats using a new microsurgical technique (20). None of the diabetic rat isograft recipients de- veloped mean blood glucose levels greater than 200 mg percent. Diabetic control lev els were 300 to 400 mg percent; normal mean values were less than 150 mg percent. However, no sur'vival statistics were reported. It is evident that such statistics are presently unacceptable if human transplant:ition is to be considered. There are a number of complications attendant to the transplantation of the pancreas in vascular continuity with the recipient which contribute to these disappointi ng re- ° sults: A. Exocrine Function. The exocrine portion of the pancreas comprises well over 90 )percent of the gland by weight. Although the acinar tissue does not contribute factors thought to' in- fluence significantly the course of diabetes in the recipient, secretions from the exocrine: pancreas pose formidable technical problems. Early attempts to tie off the pancreatic ducit at the time of transplantation led to a high incidence of pancreatitis in the graft (15,20). The exocrine secretion problem has been handled in a variety of ways: 1. Transplantation of the pancreas with a segment of duodenum as a drainage condut (14). The duodenum is anastomosed with the recipient small bowel or brought to the skin. Stich procedures have reduced the incidence of pancreatitis although the duodenum has proved to be particularly susceptible to rejection, and disruption of the duodenal stump is frequently described. 2. DL Ethionine has been employed in an attempt to suppress acinar activity of the graft without significant salutory effect (18). 3. Donor pancreatic duct ligation six weeks prior to grafting has been shown to cause fibrosi s of the acinar tissue and result in a decrease in exocrine secretion (18). Using this technique Reemtsma et al. found that they could increase the percentage of canine grafts which function but could not increase the length of graft survival. 4. Four hundred to 5,000 rad roentgen therapy will cause a decrease in the exocrine function of the gland as a result of selective injury to the acinar tissue. Use of this technic Transplantation of Insulin Secreting Tissues 281 in conjunction with previous c'uct ligation was felt to reduce the incidence of complica- tions secondary to exocrine paricreas function in dogs (15). B. Tissue Incompatiblity-Rejection.. Various immunosuppression regimen have been employed in an effort to increase the duration of transplanted allograft pancreatic function. Anti- lymphocyte serum had a modest effect in one series, increasing average graft survival from 10 to 15 days in dogs (15). Sixmercaptopurine wa's ineffective ‘in increasing functional survival of transplants in dogs. Azothiaprine increased functional survival from 5 to 15 days in dogs (18). Azothiaprine and postoperative radiation together increased survival from 5 to 20 days. To date, no immunosuppressive regimen has been discovered which consistently provides long-term protection of pancreatic allografts in experimental animals. C. Thrombosis. One of the most common causes of pancreatic graft failure has been vascular thrombosis (16,17,22,23). Although various ana:stomotic tricks have been proposed, even recent reports of pancreas transplantation have emphasized this continuing problem. Merkel et al. did, however, demonstrate a decreased incidence of botith arterial and venous thrombosis when the pan- creatic graft was interposed into the iliac circulation so that blood going to and from the leg must pass through the graft (15). D. Miscellaneous. Small bowel obstruction, :intusseception, and infection (commonly pul- monary or operative site) accounted for other postoperative deaths in pancreatic transplant recipients (24). Considerable emphasis has been made concerning the manner in which the pancreatic arterial and venous connections have been made. Physiologically, the normal pancreas is perfused with arterial blood from the aorta, and its venous drainage® contributes to the portal vein. Thus, pancreatic vascular effluent first passes through the liver before reaching the systemic circula- tion. The circulating arterial level of insulin secret agogues (glucose, amino acids, fatty acids, etc.) determine insulin release from the cells. Released insulin first reaches the liver, where a large percentage of the hormone is usually cleared. Tire physiological significance of hepatic insulin clearance and the effect of insulin on the liver is not clearly understood. Pancreatic grafts may be inserted either heterotopica lly or orthotopically. In the hetero- topic graft pancreatic venous blood drains into a systemic vein and therefore by-passes the liver. One would anticipate such a circulatory arrangement might yield abnormal systemic insulin and glucagon levels. Orthotopic grafts drain into the porta.l system. Hence the delivery of insulin and glucagon to the liver is physiologic. In an inteitesting study in pigs, Sells et al. (24) compared some metabolic effects of orthotopic and heterotopic transplants. Orthotopic graft recipients had normal glucose tolerance curves and raised plasma insulin levels. The glucose tolerance curves of the heterotopic graft recipients demonstrated significant hypoglycemia and marked hyperinsulinemia. The hyperinsulinemia in heterotopi(: transplants may be due to diver- sion of pancreatic venous effluent from the liver. The more mode st increase in insulin levels in the orthotopic transplant may reflect pancreatic denervation. In any case, one might argue a priori, that orthotopic pancre:itic transplantation much more closely approximates the normal physiologic arrangement in animals. It has been stated that one major drawback to the administration of exogenous insulin in clinical settings is that the hormone is delivered into the systemic circulation, which does not mimic the normal physiologic situation. 282 Diabetes Mellitus TRANSPLANTATION OF PANCREAS IN MAN Despite the relatively disappointing results of animal experiments, a small number of pancreas transplants have been attempted in man. Lillehei''s group in Minneapolis first trans- planted two patients in 1966. The gland with associated duodenum was placed in the iliac fossa and the vascular anastomoses were heterotopic in nature. The first patient graft survived six days, the second functioned for five months (25). In all, 36 human pancreas allografts have been done in the world. Four of these patients went 10, 12, 22, and 30 months with functioning grafts (26). One patient who received a cadaver pancreas in June 1972 remains free of the need for exogenous insulin. In addition, one patient died of a.ccidental causes ten months after trans- plantation with a functioning gland and no histological evidence of rejection at necropsy (29). Indications for pancreas grafting in man have been severe diabetic complications such as terminal nephropathy, retinopathy, neuropathy, and coronary :irtery disease. Contributing to graft failure were factors similar to those reported in experimental animals. They have been handled in the following manner: A. Exocrine Function 1. Duct ligation. Leakage from the pancreas for which no exocrine drainage route had been provided has necessitated reoperation iri the two patients receiving such a graft (27). Gliedman concluded that duct ligation is ill-advised. 2. Drainage of exocrine secretion by duodeé:nal conduit. Drainage of exocrine secretion by anastomosing the duodenum to the recipient's small intestine in a Roux en Y fashion uas been advocated. Unfortunately, both stump leaks and duodenal rejection have been com- monly described (28). It appears that the duodenum is more sensitive to rejection than the pancreas (28). In at least two instances grafts had to be removed because of duo- denal rejection even though the pancreas was functioning and appeared normal. 3. Pancreatic duct anastomosis to uret:er. Gleidmen et al. have advocated a different pan- creatic drainage technique in order to obviate those complications attendant to pancreato-duodenal transplantation (22). The utilization of direct uretero-pancreatic duct anastomosis has provided exocrine drainage, eliminated the need for co-transplantation of the duodenum and allowed the procedure to be performed outside the peritoneal cavity. In addition, pancreatic exocririe status can be assayed directly by measuring the urinary amylase concentration. In a small series of patients, this approach seems to have de- creased the incidence of compslications attributable to exocrine function of the pancreas. B. Rejection. Many of the patients receiving pancreas transplants have had histories of renal failure secondary to diabetic 'aephropathy. For this reason most pancreas transplant recipi- ents have received coincident renal allografts. Although the clincal experience is very limited and the numbers involved are small, concern that the pancreatic graft in some way elicits a more potent rejection (either to itself or to the renal allograft) has recently been voiced (27). Accordingly, it has been proposed that renal and pancreatic allografting be performed asynchro- nously, i.e., pancreas first. Using this procedure, Gleidman has improved results in terms of both number of grafts surviving ‘and length of survival. The small number of patients in this series makes any conclusion about this phenomenon tenuous. Immunosuppression regimen commonly employed included azothiaprine, antilympocyte globulin, and prednisone. The use of large doses of steroids in the pancreatic graft recipient considerably Transplantation of Insulin Secreting Tissues 283 obscures the relationship between graft function (i.e., insulin secretion) and blood glucose levels. So-called "steroid diabetes" has been reported in a number of patients. C. Vascular Complications. Many graft recipients have been reoperated upon for sudden dis- ruption of the vascular anastomoses resulting in serious bleeding (28). In summary, then, only a small number of patients have been subjected to pancreatic trans- plantation. No patients have survived longer than three years; few have had prolonged reduction in their daily insulin requirement. Little data have been collected concerning the course of diabetic vascular disease in these patients. TRANSPLANTATION OF PANCREAS FRAGMENTS OR FETAL PANCREAS IN EXPERIMENTAL ANIMALS Various attempts have been made to avoid the complications caused by the vascular anastomoses and pancreatic exocrine secretion in whole pancreas transplants. An early approach was the transplantation of fetal rat pancreas fragments. Two observations made by Gonet and Renold (31) were encouraging. First, the exocrine portion of the transplanted pancreas was noted to degenerate when the minced fetal rat pancreas was placed in the testes. Second, approximately 20 percent of the homografts (using inbred Wistar rats) demonstrated histologic evidence of endocrine pro- liferation. Furthermore, immunoreactive insulin could be extracted from the grafts. These testicular grafts were thought to represent a greater tissue volume and hence the delivery of a greater endocrine cell dose to the recipient than previously described grafts which had been placed in the hamster cheek pouch or in the anterior chamber of the eye (32,33). In approxi- mately 20 percent of the grafts to alloxan diabetic recipient rats, Gonet and Renold (31) noted a prolonged decrease in glycosuria and blood glucose. Interestingly, these same animals had the best preserved islet population when examined histologically. Further description of transplanted fetal pancreas histology was reported by Hegre et al. Portions of pancreas from fetal rats with normal or alloxan diabetic mothers were grown on organ culture for four days prior to transplanta- tion to one of two different sites in the mother, into the anterior chamber of the eye or under the capsule of the kidney. The grafts were examined four and ten days after transplantation. Following transplantation into normal mothers, the PB cells were noted to be granulated in both recipient sites. The tissue transplanted into diabetic mothers exhibited degranulation of the 8 cells. These data were interpreted as demonstrating that B cells from the fetal rat pancreas were capable of surviving for four days in organ culture and were able to maintain their func- tional integrity. Degranulation of the cells when transplanted into the diabetic mother, indi- cated that the cells were responding to hyperglycemia. Another interesting histological observation has been reported by Coupland (35). Not only did transplanted fetal pancreas fragments undergo a degeneration of exocrine cells, but activa- tion of the duct epithelial cells occurs. These cells show mitosis and form new islets of Langerhans. A similar observation has been made by Hultquist (36) when fragments from rat pancreas with a previously ligated pancreatic duct was transplanted into the anterior chamber of the eye. Despite numerous reports detailing the histological evidence of the function of pancreas fragment grafts to the eye, cheek pouch, or beneath the kidney capsule, few authors have addressed themselves to assessing the physiologic response of a diabetic animal to such a graft. Brown et al. (37), however, have recently examined the success achieved in transplanting fetal 284 Diabetes Mellitus rat pancreata into diabetic syngenic recipients. Using a streptozotocin-induced diabetic re- cipient Lewis rat, this group placed two to three pancreata of fetal age 15 to 18% days beneath the kidney capsule. The rats were treated with insulin for eight days post-transplant. Success of the graft was evaluated by urine glucose, urine volume, and serum glucose after insulin therapy had been discontinued. Sixty-four percent of the grafts resulted in complete or partial ameloria- tion of the diabetic state for up to 165 days. When the transplants were removed at 42 days from one group of rats, urine glucose and volumes rapidly increased, as did serum glucose. In- terestingly, in 10 to 20 percent of the animals, diabetes did not return when the graft was ex- cised, emphasizing the importance of adequate controls in judging the effects of "hormonal transplants (38). Histological examination of the graft revealed a multilobed organ beneath the kidney capsule comprised of mostly adipose tissue. No exocrine tissue was seen, and B cells appeared normal on electron micrographs. Unfortunately, no evidence was presented which indi- cated the response of these animals to glucose tolerance testing. Information concerning circu- lating insulin or glucagon levels was not reported. There have been only scattered reports of transplantation of fragments of pancreas in man. The most recent attempt was described in 1970 (39). A portion of a B cell tumor removed from a middle-aged woman was transplanted beneath the fascia lata in a 17-year-old juvenile diabetic. Preoperatively, the recipient had required approximately 300 units of insulin a day. One week postoperatively, the patient was aglycosuric for the first time. He was maintained on decreasing amounts of insulin, immuran, and a steroid preparation. At three months the patient required 200U of insulin daily. At that time, biopsy of the graft demonstrated granulated B cells. Nine months after grafting, a similar biopsy revealed no B cells. Despite this finding, the patient only required 125U of insulin per day. A decreasing insulin requirement in the face of histologic evidence that the graft was not functioning implies that insulin requirement is an imprecise parameter for estimating functioning B cell mass. Furthermore, it is difficult to conclude that this transplant effectively altered the course of the patient's diabetes. ISOLATION OF ISLETS The major stumbling blocks in the transplantation efforts described above are all related to an awkward anatomical reality: Pancreatic islets are distributed throughout a gland the volume of which is many times greater than that of the islets themselves. Most of the complications plaguing transplantation efforts are, then, direct corollaries of the fact that in order to trans- plant functioning endocrine tissue, which amounts to 2 percent of the pancreatic tissue mass, the remaining 98 percent, the troublesome and metabolically unimportant exocrine tissue, must likewise be foisted upon the recipient. Although minced fetal pancreas tissue had been employed in an attempt to increase the relative concentration of insulin secreting tissue in the graft, the obvious objective was to transplant islets which had been isolated from all acinar tissue. Furthermore, since islets had appeared relatively undamaged in rejected whole organs (40) it was hoped that they might be relatively protected from the ravages of tissue rejection. The hope for isolating viable intact islets of Langerhans from the pancreas lay dormat until Moskalewski (41) described collagenase digestion of the pancreas, and Lacy and Kostianovsky per- fected the method in 1967 (42). This more advanced technique entailed distention of the pancreas by the injection of Hanks solution under pressure into the common bile duct. The pancreas was then excised, finely minced, and digested with collagenase. Individual islets were then readily Transplantation of Insulin Secreting Tissues 285 visible in the dissecting microscope, and they were collected with a glass loop. Islets isolated in such a fashion appeared intact histologically and responded to insulin secretagogues. A more sensitive method for in vitro assessment of islet functional integrity was reported by Lacy et al. in 1972 (43). This technique is commonly referred to as islet perifusion. The isolated islets are placed on a millipore filter and placed in a small millipore chamber. A balanced Krebs ringers bicarbonate solution is then perfused through the chamber by a Harvard infusion pump. The effluent from the chamber which has bathed the islets is then assayed for insulin. When high concentrations of glucose are added to the perifusion medium, a rapid release of insulin occurs. The release of insulin is characteristically biphasic, there being a sharp first peak of insulin released immediately and a prolonged second phase of release which may last for many hours if the islets are continually stimulated with high concentrations of glucose (Fig. 1). This technique has proved to be very sensitive and islet perifusion has been a critical test for assessment of islet viability. In addition, this method has become a potent basic science investigative tool. One further refinement of islet isolation has increased the yield of islets, decreased the time required to obtain them, and has led to a burgeoning of isolated islet transplantation efforts. This refinement consisted of using a density gradient to separate islets from acinar debris after collagenase digestion. Originally, a sucrose density gradient was pro- posed (42). Lindall et al. suggested a discontinuous gradient of different concentrations of ficoll, a polymer of sucrose (44). This technique was explored by Scharp et al. who found that increased yields of viable islets were obtained with the method (45). They emphasized the im- portance of dialysing the ficoll before exposing islets to the polymer. These techniques have allowed the isolation of functionally intact islets of Langerhans from rat pancreata in relatively large numbers. Once the isolation of islets of Langerhans had been described and these islets had been demonstrated to be intact functionally, the first attempts at transplantation of isolated islet followed soon afterward. The subsequent flurry of transplantation projects has generated a great deal of excitement among researchers, clinicians, and the public alike. TRANSPLANTATION OF ISOLATED ISLETS IN EXPERIMENTAL ANIMALS In an early report describing the transplantation of isolated islets, Reemstsma took ad- vantage of the relative ease with which piscine islets could be obtained free of acinar tissue (46). Thus, fish islets were transplanted into rats which had been made diabetic by the ad- ministration of streptozotocin. A variety of methods of islet implantation were employed: 1) islets in millipore chambers placed in the peritoneal cavity, 2) islets implanted intra- muscularly, and 3) islets inserted into the anterior chamber of the eye. Although a decrease in blood sugar in the diabetic recipient rats was described, this effect was short-lived, generally not lasting for more than four days. The first major attempt at transplantation of isolated islets from one animal to another of the same species was reported by Ballinger and Lacy in 1972 (47). Using inbred strains of rats, these experiments were not subject to the complications of tissue immunorejection. The recipient animals had been rendered diabetic with streptozotocin. Blood glucose, urine volume, urine glucose, and weight were monitored in both control and experimental animals. The transplantation of 400 to 600 isolated pancreatic islets into the peritoneal cavity or into the thigh muscle of the recipient resulted in a significant long-term reduction of hyperglycemia, polyuria, and glycosuria, and a restoration of weight gain. 286 Diabetes Mellitus Glucose, mg/ml v 03 y 30 Cy ND I INSULIN uo Units/Islet/min ¥ 1 Obs 1 1 1 Sens 1 mal 4 l ok 1 1 20 30 40 50 60 70 80 90 MINUTES FIGURE 1. Biphasic pattern of insulin secretion fol- lowing stimulation of perifused rat islets with glucose (3.0 mg/ml). Vertical lines represent S.E.M.. From Lacy, Walker, and Fink. Diabetes 21:987-98, 1972. Re- printed with permission from Diabetes the Journal of the American Diabetes Association. Although the average urine glucose, urine volume and blood glucose levels were improved in the transplanted group, these values did not approach the normal control levels. Nonetheless, some individual animals did achieve normal control values for two months or more. Excision of islets which had been transplanted into the thigh muscle resulted in a rapid return to the diabetic state. Histologic examination of the excised islets revealed intact o and B cells with a marked degranulation of B cells; indicative of the physiological demand for insulin placed on these cells. This report also detailed preliminary allograft experiments. Islets were transplanted across a major histocompatibility barrier and the recipients received immunosuppression. The animals subjected to this protocol evidenced some amelioration of their diabetes. These experiments were the first to demonstrate that isolated islets were capable of per- manently reversing chemically induced diabetes in experimental animals, and they served to acti- vate islet transplantation interest in other laboratories. The obese hyperglycemic mouse was used as an experimental animal in a few early studies ex- amining the effects of islet transplantation. Obese mice provide an interesting model, exhibiting a syndrome characterized by obesity, elevated blood glucose levels, and markedly high circulating insulin concentrations. Isolated islets were placed in a millipore diffusion chamber and the chamber was placed in the peritoneal cavity. A chamber pore size of 0.45u presumably allows Transplantation of Insulin Secreting Tissues 287 ingress of insulin secretagogues and the egress of insulin and glucagon. When obese mice re- ceived normal mouse islets, weight gain stabilized, blood glucose levels fell, and interestingly, a dimunition of insulin levels followed (47). The mice reverted to pretransplant phenotypes when the chamber was removed. These data were interpreted as indicating that the obesity syndrome in the obese mouse was secondary to a defect in a factor elaborated by pancreatic islets which was capable of passing through the millipore diffusion chamber. Gates et al. (50) extended these studies, observing normalization of obese mouse blood glucose and insulin levels for up to 10 weeks. Oral glucose tolerance tests in transplanted animals approached the response seen in normals. Leonard et al. have performed a series of transplants in rats, using neonatal pancreas which has been finely minced and digested with collagenase although the islets were not separated from remaining acinar tissue (53). Pancreata excised from neonatal rats on days two and a half to four and a half postpartum, had the lowest exocrine enzyme concentration and the highest rela- tive insulin content. Tissue prepared from donors of that age was transplanted into the peri- toneal cavity of either semi-inbred (homologous) or highly inbred (isologous) diabetic recipients. Ninety-six percent of the homologous rats had an amelioration of their diabetes for 3 to 13 days, with an average duration of 10 days. Moderately diabetic isologous recipients, on the other hand, all experienced normoglycemia, persisting in some up to five months. SITE OF IMPLANTATION There are several important theoretical and practical considerations involved in the selec- tion of an appropriate site for islet administration in the diabetic animal, especially man. Ideally, the site should be accessible with only a minor operative procedure. The organ or tissue into which the islets are placed should be expendable so that should an untoward tissue reaction occur, the graft can be removed. Finally, although the significance of the peculiar vascular arrangement of the pancreas (arterial input, portal vein effluent) is not fully appreciated, it is likely that such a configuration is of physiologic import. The profound effects of insulin on the liver and the remarkable ability of that organ to clear the hormone, imply that the meta- bolic relationships between normal pancreatic islets and the liver are intimate. The importance of the site of islet implantation was explored by Kemp et al. (53). They found that 600 to 850 islets placed subcutaneously had no significant effect on the urine glucose, urine volume, or blood sugar of diabetic isologous rat recipients. A similar number of islets transplanted into the peritoneal cavity resulted in amelioration of the diabetic state, but none of the parameters examined reverted to normal values. When an equal number of islets were injected into the portal vein, the results were more dramatic. The diabetic animals achieved normal urine volumes and blood glucose levels (Figs. 2 and 3). Histologic examination of the recipient liver revealed intact, vascularized islets lodged in the terminal portal tracts. The selective ability of islets injected into the portal vein to reverse chemically induced diabetes in the rat in these studies is a provocative observation. It is clear that whatever factors operate in this experimental situation to allow islets lodged in the portal tract to revert the diabetic animal to normal must be investigated further. Questions deserving investigation include: 1) Do a greater percentage of islets survive in the liver than other sites? 2) Is B cell replication more pronounced in the liver? 3) Do portal tract administered islets release more insulin (less glucagon) than islets transplanted to the peritoneal cavity or sub- cutaneous tissues? 4) What, exactly, is the significance of insulin delivered in the portal 288 Diabetes Mellitus circulation versus that in the systemic circulation? 5) How does the liver respond to islets lodged in its parenchyma? 6) How, if at all, are islets innervated in the various transplantation sites? 7) What is the exact nature of the transplanted islet's blood supply once lodged in the liver? and 8) What effect does portal blood have on insulin secretion? Presumably portal blood should be rich in absorbed foodstuffs and thus present higher levels of insulin secretogogues to ‘the islets lodged in the liver. Urine Glucose Diabetic Control Group (10) gait unr oveat o Subcutaneous Group (5) @® I ~~ Streptozotocin ~=———— |mplantation oS : 5 Intraperitoneal Group (5) 2+ Lod Ld 1 4 4 1 J 1 | Intraportal Group (5) 0) 20 30 40 50 60 70 80 Time In Days FIGURE 2. Effect of portal vein, subcutaneous, and intraperitoneal implantation of pancreatic islets on the urine glucose of diabetic rats. Diabetic controls. From Kemp, Knight, Scharp, Ballinger, and Lacy. Diabetologia 9:486-491, 1973. Reprinted with permission from Springer-Verlag. IMMUNOSUPPRESSION There have been conflicting reports on the relative ability of endocrine tissue to provoke transplantation immunorejection. Reckard et al. (51) examined the fate of isolated islets trans- planted into rats which had a major histocompatability difference from the donors. Using a streptozotocon-induced diabetic recipient and administering 600 to 1200 islets into the peritoneal cavity, these investigators found that the islets did indeed provoke rejection. In fact, several other interesting observations were made: 1. Islet homografts functioned only one to three days when donor and recipient differed by a strong histocompatability (Fischer to ACI). The period of functional survival was not lengthened by a course of immunosuppression (antilymphocyteserum). 2. Islet homografts compatible at the AgB locus (ACI to DA) resulted in normoglycemia of the recipient for a mean of twelve days. Intriguingly, antilymplocyte serum had a pronounced effect on these animals, extending the median functional islet survival to 30.5 days. 3. To determine if the recipients of homologous islets had become sensitized to donor tissue, six Fischer rat recipients of 800 to 1200 AgB incompatible Lewis islets were challenged with donor skin grafts. These grafts were rejected in an accelerated manner, indicating that homologous islets did indeed stimulate the host immune system. By Transplantation of Insulin Secreting Tissues 289 Blood Glucose Diabetic Control Group (10) Lif ssbcsone Group (5) SS o 3 ~=———— Streptozotocin E . Intraperitoneal 3 NS rgperien = 300 Se Q QQ o 200 tS 100 Int tal gwd bell i pe a dt hdd 0 10 20 30 40 50 0 70 80 Time In Days FIGURE 3. Effect of portal vein, subcutaneous, and intraperitoneal implantation of pancreatic islets on the blood glucose of diabetic rats. Diabetic controls. From Kemp, Knight, Scharp, Ballinger, and Lacy. Diabetologia 9:486-491, 1973. Reprinted with permission from Springer-Verlag. contrasting the period required for tissue rejection to become manifest, the authors concluded that islet tissue is at least as antigenic as skin or heart. It should be noted, however, that there was no attempt to control the dose of islets administered in comparison to amount of skin or heart transplanted. Furthermore, the animals in the study were selected without regard to sex. It appears that there may indeed be a difference between males and females in regard to their ability to mount an immuno- rejection response. Finally, the authors have employed a physiologic end point to determine islet viability (i.e., the ability to secrete insulin) whereas they have used an anatomic end point to assess skin graft survival. EFFECT OF ISLET TRANSPLANTATION ON THE COMPLICATIONS OF DIABETES Some preliminary evidence concerning the effect of islet transplantation on the secondary complications of diabetes has been gathered. Sutherland et al. (54) have described renal glomerular lesions in rats six months after induction of diabetes characterized by immuno- glubulin and complement deposition in the glomerular measangium. This is followed by mesangial matrix thickening. In diabetic rats which had received intra-peritoneal transplantation of isologous neonatal pancreatic islet tissue, plasma glucose levels were significantly lower and several serial biopsies showed progressive decrease in mesangial immunoflourescent staining for IgG, IgM, and B;C four to nine weeks after transplant only traces of the above could be detected. Mesangial matrix thickening was arrested or actually was reduced. The relationship between mesangial deposition of complement and immoglobulin and streptozotocin induced diabetes is not 290 Diabetes Mellitus clear. The alteration of these lesions by islet transplantation, although intriguing, cannot be interpreted easily. TRANSPLANTATION OF ISOLATED ISLETS IN OTHER ANIMALS Islet transplantation in man will pose two problems not addressed by the model defined in the rat. The pancreas in the human is compact and fibrous as compared to the readily dis- tensible gland in the rat. In addition, inbred strains of humans do not exist. To examine these problems, preliminary work has been reported in animal models thought to be more relevant to the problems posed by human islet transplantation. The isolation of islets and the transplantation of same in both pig and monkey have been described. Although the more compact pancreas of the pig and the monkey pose formidable islet isolation problems, some of the challenges have been met. Nonetheless, the in vitro assay of these isolated islets had not been as impressive as those described for rodents. A great dedl of work remains to be done. Transplantation of these islet preparations has also been described. Again, the results are less impressive than the re- sults described in the rat. Here, too, many more experiments will be required before these higher animal models of islet transplantation are fully developed. TRANSPLANTATION OF ISOLATED ISLETS IN MAN Human islets have been successfully isolated from recently expired donors or from operative specimens. The compact fibrous nature of the human pancreas makes this feat much more difficult than isolating islets from the diffuse, easily distended rat pancreas. Human isolated islets have been shown to release insulin when exposed to high concentrations of glucose in a perifusion system. Isolated islets have been transplanted into the peritoneal cavity of at least one . diabetic patient by the Minnesota group (57). Unfortunately, no long-term salutory effect was documented in the one patient described. It is clear that a great deal of intensive investigation needs to be done before routine islet transplantation in man is a clinical reality. A number of the questions remaining to be answered might well be worked out in animal models before further clinical experimentation takes place. A list of such questions might include: 1. What number of islets needs to be transplanted to achieve normalization of either spontaneous or chemically induced diabetes? 2. Which recipient site of transplantation maximizes the chance of islet survival and replication, yet minimizes the likelihood of tissue rejection? 3. What is the most appropriate assay of islet viability, such that only live functioning islets will be transplanted? 4. What various types of islet preservation will prove feasible? It is clear that islet tissue may become available at a site far removed from the prospective recipient. Methods of storage and preservation must be developed. Furthermore, pooled islets from more than one donor may be required to successfully transplant one recipient. Two types of preservation are now under active investigation: a) Cryopreservation. The ability to store certain blood elements at very low tempera- tures has stimulated investigation of cold storage of isolated islets. Islets have been stored in Hanks solution at 4°C for up to 48 hours with some success (58). Knight has preserved islets in DMSO at -180°C for up to 14 days with maintenance of islet functional integrity as assayed Transplantation of Insulin Secreting Tissues 291 both by perifusion and transplantation. The appropriate storage media, methods of cooling and thawing must be further explored, as successful cryopreservation will most likely be a cornerstone of clinically applicable islet transplantation. b) Tissue Culture. Tissue culture has been the other major approach to islet preserva- tion. This is for several reasons. First, it is hoped that islets might be stored for long periods of time without loss of functional integrity or dimunition of numbers. Second, some optimism has been expressed that islets on tissue culture may replicate, hence increasing the original islet yield from a donor pancreas. Further, it had been postulated that tissue maintained in organ culture might lose those antigenic determinants which provoke host versus graft tissue rejection (the so-called Summerlin hypothesis). Some preliminary data have been collected which bear upon the above postulates. Islets have been maintained in tissue culture successfully by many groups (38). They maintain their ability to synthesize proinsulin and insulin and will release the latter in response to glucose stimulation. Acinar contamination of isolated islets seems not to survive in tissue culture. Although in vitro culture of the whole fetal rat pancreas by Hegre et al. (38) demonstrates a relative increase in islet volume in comparison to exocrine tissue, conclusions about the ability of isolated islets to replicate in cell culture are difficult to make. Some of the most elegant work on culturing of islet tissue has been described by Chick (38). Monolayer cultures of dis- persed cells from the neonatal rat pancreas have been carried out. Cells cultured in such a fashion respond to glucose by releasing insulin and they have maintained this ability for a period of one year. Furthermore, it is thought that these cells actively replicate, although this process is ultimately inhibited by fibroblast overgrowth. 5. What type of immunosuppression will be best suited for discouraging immunologic response to transplanted islets? Although a course of antilymphocytic globulin has proved to be the most efficacious immunosuppression regimen in preliminary allograft studies in rodents (51), little systematic evaluation of immunosuppression protocols has been described. Although transplantation of insulin producing tissues as a treatment of diabetes mellitus was proposed almost three quarters of a century ago, the hint that such an approach might be clinically relevant has come within the past decade. Recently, experimental isolated islet trans- plantation has developed rapidly, making it the most likely approach to prove useful for human diabetics. Yet despite the rapid growth of accumulated information about isolated islet trans- plantation, there remains a great body of knowledge still to be collected. The scientific com- munity has learned that premature emphasis of transplantation of certain organs ultimately retards progress and takes lives. The decision to intervene in the course of a diabetic patient will generate even greater agony than the decision to transplant for a failing heart. Furthermore, it seems apparent that insulin-producing tissue will be subject to all the manifestations of immuno- rejection seen with other transplanted organs. Yet the notion of definitive treatment of the diabetic patient is a glamorous and very im- portant goal. The rationale for the transplantation of insulin secreting tissues is to see whether functioning normal 8 cells will prevent or arrest the progression of the complications of diabetes. It is imperative that detailed studies be carried out on the tissue transplanted and the effect it has upon the recipients. Longitudinal studies designed to assess the efficacy of the transplant over a period of years will be central to our understanding of this approach to the treatment of 292 Diabetes Mellitus diabetes. A final benefit of exploring islet transplantation is the great bulk of basic science information generated. This information and experience may ultimately prove to be the most meaningful of all. REFERENCES 1. Marble, A, P White, R Bradley, and L Krall 1971. Joslin's diabetes mellitus, p 7. Philadelphia, Lea and Febiger. 2. University Group Diabetes Program 1970. Diabetes 19 (Suppl. 2):747-830. 3. Prout T, G Knatterud, C Meinert, and C Klint 1972. Diabetes 21:1035-1040. 4. Williamson, T, N Vogler, and C Kilo 1969. Diabetes 18:567-578. 5. Kilo C, N Vogler, and T Williamson 1972. Diabetes 21:881-905. 6. Williamson T, N Vogler, and C Kilo. Med Clin North American 55:847-860. 7. Becker, D, and M Miller 1960. New England J Med 263:367. 8. MacDonald, M and J Ireland 1964. Aetiology of diabetes and its complications. M. Cameron and M. O'Conner (eds.) Ciba Foundation on Endocrinology. Vol 15, Little, Brown § Co. 301-314. 9. Cerasi, E, and R Luft 1970. Pathogenesis of diabetes melliutus. Cerasi, E, and R Luft (eds.) Nobel Symposium 13. Almqvist and Wiksell, 17-43. 10. Felig, P 1971. Med Clin North Am 55:821-834. 11. von Mering, J. and O Minkowski 1889. Arch Exper Path and Pharm 26:371-374. 12. Hedon, I 1892. Arch de Physiol Norm et Path 5:617-628. 13. Ssobolew, LW 1902. Arch Path Anat Klin Med 168:91. 14. Largiader, F, G Lyons, F Hidalgo, R Dietzman, R Lillehei 1966. Am Jour Surg 113:70. 15. Merkel F, W Kelly, F Goetz, and J Maney 1968. Surgery 63:291-297. 16. Pemberton, LB, and W Manax 1971. Sur Gyn Obs 75-80. 17. Aquino, C, J Ruiz, LS Schultz, and RC Lillehei 1973. Am J Surg 125:240-244. 18. Reemtsma, K, N Giraldo, DA Depp, and E Eichwald 1968. Ann Surg 168:436-446. 19. DeJode, LR, and M Howard 1962. Surg Gyn Obs 553-558. 20. Kelly, W, RC Lillehei, FK Merkel, Y Idezuki, and F Goetz 1967. Surgery 61:827. 21. Archambeau, J, M Greim, and P Harper 1966. Radiat Res 28:243. 22. DeJode, LR, and JM Howard 1966. Brit J Surg 53:364. 23. Chaja, A, H Appert, and J Howard 1966. Arch Surg 93:953. 24. Sells, R, R Calne, V Hadjiyanakis, and V Marshall 1972. Brit Med Journ 3:678-681. 25. Kelly, M, R Lillehei, and F Merkel 1962. Surgery 61:827-837. 26. ACS/NIH 1973. Organ transplant registry. Jour Am Med Assoc 226:1216. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. Transplantation of Insulin Secreting Tissues 293 Gleidman, M, M Gold, J Whittaker, H Rifkin, R Soberman, S Free, V Tellis, and F Veith 1973. Surgery 74:171-180. Lillehei R, RL Simmons, J Najarian, C Kjellstrand and F Goetz 1971. Transpl Proc 3:318-324. J Connolly, D Martin, T Steinberg, G Gwinup, A Gozzangia, and R Bartlett. Arch Surg 106: 489-494. Lee, S, J Chandler, R Krubel, N Nakaji, H Rosen, and M Orloff 1972. Surgical Forum xxii: 75-77. ] Gonet A, and A Renold 1965. Diabetologia 1:91-96. House R, C Burton, H Cooper, and E Anderson 1958. Endocrinology 63:389. Coupland, RE 1957. Nature 179:51. Hegre, O, L Wells, and A Lazarow 1970. Diabetes 19:906:915. Coupland R 1960. Jour Endocrinology 20:69. Hultquist, C, and J Upsala 1972. Med Sci 77:8-18. Brown, J, I Molnar, W Clark, and Y Mullen 1974. Science 1377-1379. Goetz, FC 1974. Metabolism 23:875-884. Urca, I, I Kott, and A Lorkan 1970. Diabetes 19:182-185. Lillehei, R, R Simmons, J Najarian, J Weil, H Uchida, J Ruiz, C Kjellstrand, and F Goetz 1970. Ann Surg 172:405. Moskalawski S 1965. Endocrinology 5:432. Lacy, P, and M Kostianovsky 1967. Diabetes 16:35. Lacy, P, M Walker, and C Fink 1972. Diabetes 21:987-988. Lindall, A, M Steffes, and R Sorenson 1969. 85:218. Sharp, D, C Kemp, M Knight, W Ballinger, and P Lacy 1973. Transplantation 16:686-689. Reemtsma K 1970. Transplantation Proc 2:513-515. Ballinger, W, and P Lacy 1972. Surgery 72:175-186. Younoszai, R, R Sorenson, and A Lindall 1970. Diabetes (Suppl) 19:406. Strautz, R, 1970. Diabetologia 6:306-312. Gates, R, M Hunt, R Smith, and N Lazarus 1972. Lancet 567-570. Reckord, C, M Zinger, and C Barder 1973. Surgery 74:91-99. Kemp, C, M Knight, D Scharp, W Ballinger, and P Lacy 1973. Diabetologia 9:486-491. Leonard, R, A Lazarow, and O Hegre 1973. 22:413-428. Sutherland, D, M Steffes, S Manor, and J Najarian 1974. Surgical Forum XXV:309-311. Sutherland, D, M Steffes, G Bauer, D McManus, B Noe, and J Najarian 1974. Surg Research 16:102-111. Scharp, D, J Murphy, W Newton, W Ballinger, and P Lacy 1975. Surgery 77:100-105. 294 Diabetes Mellitus 57. 58. 59. Najarian, J, 1974. Reported at annual diabetes association twelfth research symposium. Minneapolis, Minnesota, October 25-26, 1974. Knight, M., D Scharp, C Kemp, W Ballinger, and P Lacy 1973. Cryobiology 10:89-90. Scharp, D, C Kemp, M Knight, J Murphy, W Newton, W Ballinger, and P Lacy 1974. (Suppl) 23:359. Diabetes 22. DIET AND DIABETES MELLITUS Ronald K. Kalkhoff Elsewhere in this monograph a variety of experimental and clinical evidence has been assembled that stresses the importance of controlled metabolism in the diabetic patient. Obvi- ously, a most important means of good clinical management is strict adherence to an acceptable dietary regimen. In the following discussion, the nutritional aspects of diabetes mellitus are related to its incidence, control, and potential complications. Controversies in various research areas are delineated, and suggestions are made regarding the need for additional studies. Final- ly, the practical problems of diet and diabetes that confront physician and patient are emphasized and possible solutions to them are offered. THE IMPACT OF NUTRITION ON DIABETES PREVALENCE There is abundant evidence suggesting that average caloric intakes of a given population have a substantial influence on the prevalence of diabetes mellitus. Thus, in underdeveloped countries where total caloric intake is reduced below the average of more affluent nations of the world, the incidence of the disease has been less in many instances (20). Extremes of malnutrition provide greater support for this point. During the latter portions of World War II, countries devastated by war developed food shortages even to the point of death from starvation. While these events were very unfortunate, epidemiologists recognized a concom- itant fall in the incidence of clinically symptomatic, diagnosable diabetes mellitus. In Japan, for example, newly discovered cases of diabetes in one major clinic were reduced in number nearly threefold (7). During the next 10 years after the war's conclusion, an increase in food supplies correlated well with a proportional increase in diabetes prevalence (Fig. 1). At the other extreme of malnutrition, i.e., obesity, one sees the opposite effects on dia- betes prevalence occurring. Approximately 50 percent of obese, middle-aged adults exhibit some form of carbohydrate intolerance, and about 40 percent of all adult-onset diabetic subject are obese. The higher incidence of diabetes among more affluent populations generally has been shown to relate best to the prevalence of overweight, obese states. All of these observations suggest that overnutrition unmasks the diabetic syndrome, whereas normal or restricted food intake has some protective effect on diabetes-prone individuals. Un- fortunately, these observations have gone unheeded, since the prevalence of obesity is steadily rising in the United States and other advanced countries each year (18). INFLUENCE OF DIET COMPOSITION ON DIABETES PREVALENCE Although it is generally agreed that caloric excess and obesity increase the prevalence of diabetes, there is much controversy about the influence of dietary composition on this parameter. Himsworth (9) proposed that diets high in fat content were chiefly responsible for the higher incidence of diabetes in certain populations. However, the American Eskimo, who subsists on a high fat diet, has a low incidence of the disease. West and Kalbfleisch (20) have shown that in rural Uruguay, where a high fat-protein diet is common, the incidence of diabetes is quite low, whereas 295 296 Diabetes Mellitus 2.0 Ratio of Type 1 7. Diabetics to Type 2 10 Diabetics FIGURE 1. Fluctuation in number of dia- betic patients, ratio of diabetics to | 7.5] Ratio of total outpatients and ratio of type 1 Diabetics to diabetics (mild and obese) to type 2 Total diabetics (severe and thin) during 1936 7.0 Outpatients = through 1956. Pacific war: Dec. 1941 ipa - to Aug. 1945. From Goto et al. 1958. Reprinted with permission from Diabetes, the Journal of the American Diabetes os [] Association. 60 S01 40r 30 20} |, Number of Diabetic Patients LL 100 0 ms 0 1455 in urban Uruguay, where composition of diets among relatives of the rural group is similar but obesity is more common, there is a higher frequency of the disease. Nutritionists also have examined the possible relationship between carbohydrate consumption and diabetes. Yudkin (21) concluded that Himsworth's statistical association of diabetes and vascular disease with the degree of fat intake did not take into account concomitant increases in dietary consumption of refined sugar in these same populations (Fig. 2). His epidemiologic data for 22 countries demonstrated a highly significant relationship between the amount of sugar con- sumed and mortality rates due to diabetes. Campbell (4) in his analysis of various ethnic groups, reached similar conclusions and particularly stressed the increased use of refined sugar as having a causative role in the dietary aggravation of diabetes. A general epidemiologic study of diet and diabetes in Central and South American countries uncovered a similar trend between sugar in- take and diabetes prevalence, but on a statistical basis, this could not be proven (20). Not all published reports favor the sucrose-diabetes relationship. Among five geographic areas in Trinidad, the incidence of diabetes was lowest in those districts with the highest annual per capita consumption of refined sugar (13). Moreover, nutritional studies of Pima Indians in Arizona, whose prevalence of diabetes is 10 to 15 times greater than the general Diet and Diabetes Mellitus 297 175 F 1 ® 150 | . - ° ns oo . 3 yy 125 4 § ro” R . . % 100 | - < ¢ ° ° RQ 5 o ° by ? ® ° oe N < Sof ’ oe . J ® . ® st, ® % - 1 Le A | L 0 25 50 75 100 125 150 175 SUGAR INTAKE ( grammes per day ) FIGURE 2. Relation between average fat in- take and average sugar intake in 41 countries. From Yudkin 1964. Reprinted with permission from Lancet. population in the United States, also do not support the view that selected nutrients profoundly influence the prevalence of this disease. Detailed dietary histories of 248 or over 80 percent of the young adult female population in this tribe compared food intake of 169 nondiabetic and 79 diabetic subjects. Although the diabetic women were significantly more obese, their total carbo- hydrate and sucrose intakes were significantly less than corresponding intakes of the nondiabetic group (14). = It would seem from these studies that the relationship between the distribution of carbo< hydrate:, fat, and protein in a given diet and the prevalence of diabetes remains highly contro- versiall and unsettled. EFFECTS OF DIET ON PREEXISTING DIABETES IN THE OBESE Clinical investigations have provided further insight into the greater prevalence of dia- betes among overweight individuals. In both the fasting and postprandial state, the maturity- onset diabetic individual has higher plasma insulin concentrations than the nonobese, diabetic subject. This suggests that insulin is less effective in controlling glucose levels when obesity is superimposed on diabetes. There are several reports demonstrating that with weight loss, carbo hydrate tolerance improves despite lower insulin concentrations in thinned obese, indicating an amelioration of the resistance to endogenous insulin and increased efficiency of hormonal action (10), (Fig. 3). [Disturbances in circulating lipids including triglyceride and cholesterol, also are fre- quently improved with weight reduction (6). It is also reported that at least a partial correla- tion b etween plasma insulin concentrations, serum glucose, and serum triglyceride levels exists in obe sity (1). The restoration of insulin to normal levels after achieving ideal body weight may al so serve to improve blood lipid as well as blood glucose abnormalities. 298 Diabetes Mellitus CONTROL E © o—o Obese Before o 50/ 150 Wat. Loss Oo S e—o Obese After 2 125} 125 Wat. Loss I 8 100+ 100 oO 3 oOo eo =75] 75 £ a O 60 120 180 240 0 60 120 180 240 7 200} 200 3 FIGURE 3. Plasma glucose and insulin 1 50} ' 150 responses and insulin-glucose (I/G) 5 concentration ratios in six obese patients 2 100t 100 before and after weight reduction and in ° ten nonobese patients. Values are mean £ | 50 + 85.E.M. Asterisk: significance of the 8 50 differences between corresponding means 5 5 in the obese group before and after weight Oo eT loss, p <0.05. Plus sign (+): signifi- = 0 6 20° 150 240 cance of the differences between corre- o sponding means of obese and control subjects si 15 by unpaired data analysis, p <0.05. From 2 Kalkhoff et al. 1971. Reprinted with per- 5 LO mission from Diabetes: The Journal of the 2 ’ American Diabetes Association. o 3 5% 5 o gE © 9 a oL8 O 60 120 180 240 0 60 120 180 240 Minutes Minutes It is also of interest that several animal models illustrate the antagonistic effect of ! obesity on diabetes. For example, in the New Zealand obese mouse, an animal that is genetically prone to obesity and carbohydrate intolerance, overeating and weight gain promote the development of diabetes, as well as elevated basal insulin concentrations in the blood. Caloric restrictions and avoidance of weight gain effectively obviate these complications. Thus, a major principle in diabetes management is prevention of obesity and aggressively treating overweight diabetic subjects with reasonable caloric restriction until ideal body weijzht is achieved. DIET AND THE NONOBESE DIABETIC SUBJECT Concepts concerning the ideal diet for diabetic patients who are nonobese are constantly changing. Because a basic problem is control of high blood sugar concentrations, it was recomi- mended for several years to restrict carbohydrate intake to a moderate extent. The formulation recommended by the Council on Foods and Nutrition in 1958 included total calories of 30-35 calories per kilogram body weight for middle-aged, nonobese adults. Calorie:s Diet and Diabetes Mellitus 299 derived from carbohydrate and fat were each 40 percent of total intake, concentrated sweets were restricted, and a substantial portion of fat calories were to be ingested in the polyunsaturated form. The remaining 20 percent of calories was to be derived from high quality protein foods to insure at least one gram of protein per kilogram body weight each day. Since carbohydrate con- stitutes 45 to 50 percent of calories consumed by Americans, this recommendation did not represent a serious departure from eating habits in the United States. However, in several countries throughout the world, particularly where rice and other grains are the main food staple, daily carbohydrate intake may greatly exceed 50 percent of total con- sumed calories. Nevertheless, diabetic diets tailored to this type of eating pattern have shown no adverse effects on control of hyperglycemia or insulin requirements even when total daily calories as carbohydrate approached 70 percent. This experience has been shared by several in- vestigators,including those in Europe and the United States and for periods of follow-up as long as 8 years. This has led a special committee of the American Diabetes Association (2) to modify recommendations concerning diabetic diets. In individual cases, this group stated, liberaliza- tion of carbohydrate intake does not appear to be contraindicated, providing appropriate caloric intake is not exceeded and other metabolic parameters of good diabetic control are not disturbed. DIETARY CARBOHYDRATE IN DIABETES MANAGEMENT There are additional observations suggesting beneficial effects of high carbohydrate diets on diabetic patients. This has been summarized recently by Brunzell and colleagues (3). In their studies, a change-over from a balanced regimen consisting of 45 percent carbohydrate, 40 percent fat, and 15 percent protein to an 8- to 10-day course of an 85 percent carbohydrate, 15 percent protein, 0 percent fat diet significantly lowered fasting and postprandial blood glucose con- centrdtions during 100 gram oral glucose tolerance tests in mild, adult diabetic subjects (Fig. 4). Fasting insulin concentrations were also lowered significantly. The authors concluded that high carbohydrate diets render insulin more efficient in the control of blood glucose concentra- tions in diabetic patients. aor - Basal E 160 = AX es... S ‘o... 2 Tog re susiEaRTeRaRRE 505m nse Hr o 140 High Carbohydrate E 9 ¥20 FIGURE 4. Mean oral glucose tolerance on basal and high 0 : carbohydrate diets. By paired comparisons, values both > i fasting and after oral glucose were significantly de- > 100 7 creased on the high carbohydrate diet. From Brunzell et al. 1971. Reprinted with permission from the New O England Journal of Medicine. E so w > a. 60 | | | | | | 0 30 60 90 120 150 180 Time (minutes) 300 Diabetes Mellitus Others have reported improved carbohydrate tolerance in normal men following treatment with high sucrose diets on a short-term basis, but longer administration of this regimen resulted in no beneficial effect as compared to a more favorable outcome with high starch diets. Still others observe that while high carbohydrate diets may improve oral 100 gram glucose tolerance tests, they do not necessarily lower blood sugar or insulin responses to a more typical mixed meal; in fact, a worsening effect may be demonstrated under these conditions. The foregoing observations lead to widely divergent conclusions concerning high carbohydrate diets in control of blood glucose homeostasis in diabetes. There are more uniform opinions re- garding low carbohydrate diets, however. It is well known, for example, that restricted carbo- hydrate diets lead to impaired oral glucose tolerance. Basal insulin secretion is also decreased. More recently it has been reported that limited carbohydrate intake may promote deterioration of glucose tolerance by impairing peripheral utilization of this fuel in tissues such as skeletal muscle. Apart from diabetogenic effects, low carbohydrate diets severely restricted in calories, as in quick weight loss schemes for obese subjects, may promote demineralization of the skeleton and increased urinary losses of calcium and other minerals. It would appear that severe limits placed on, carbohydrate intake (20 percent of total calories or less) have no important role in the dietary management of most diabetic patients today. DIETARY COMPOSITION AND DIABETIC SERUM LIPIDS Factors responsible for elevated blood lipid levels in diabetes are most complex and are re- viewed in greater detail in other portions of this monograph. In brief, the vast majority of disturbances of triglycerides are due to overproduction of this moiety by the liver and increased entry into the systemic circulation. Defective removal of triglyceride by peripheral tissues also contributes to a greatly expanded blood triglyceride pool (6). Most of these abnormalities, as mentioned earlier, are corrected by reduction of obese patients to ideal body weight and optimum control of blood glucose levels by diet alone or in combination with oral agents and in- sulin. Nevertheless, a segment of the diabetic population will continue to manifest blood lipid abnormalities despite these measures. Some of these individuals are unduly sensitive to carbohydrate; others may demonstrate sensitivity to both dietary fat and carbohydrate. These two groups have been classified into Types IV and V acquired lipid disturbances by Fredrickson and Levy (6), based on laboratory measurements of lipid-carrier proteins or lipoproteins in blood. However, to complicate this scheme further, additional research has recently shown that type IV lipoprotein disturbances are found in a heterogenous collection of patients whose sensitivity to carbohydrate and fat in the . diet is highly variable and not readily predictable by lipoprotein classification. Some do re- quire restriction of carbohydrate, others, limited fat intake, and some respond best to low fat and carbohydrate diets based on empirical observations. These data have particular relevance to the trend toward liberalizing carbohydrate intake in diabetic diets. While this manipulation may be of no consequence in the control of blood glucose concentrations, the undesirable side effect of triglyceride elevations may preclude its use in certain individuals. Kaufmann and Stein (11), in their review of the subject, note that the induction of hyper- lipemia by carbohydrate may be more contingent on the type of nutrients administered. Sucrose may have more profound effects than other types of sugars, and the triglyceride response may be Diet and Diabetes Mellitus 301 exaggerated to a greater degree when saturated, as opposed to unsaturated fats, are a part of the regimen. They also cited evidence that men and post-menopausal females more frequently manifest this phenomenon. Refined sugar has been implicated in the causation of hyperlipemia by several authors, as reviewed by Roberts (16). He also showed that omission of sucrose from diets given to men with elevated serum triglyceride concentrations effectively improves the condition in most instances (Fig. 5). . HIGH-TRIGLYCERIDE 190} Sucrose-free diet GROUP . 180 170 5 160 a2 oO © © Sucrose-free diet LOW -TRIGLYCERIDE 120 GROUP v oc = Oo oO © o 1. 1 1 1 1 1 I I ! I 1 I I 1 1 1 1 1 | 1 1 | I ! | 1 1 1 I 1 i 1 | 1 ' 1 1 FASTING SERUM -TRIELY CERIDE (mg. per /00m/.) ol o 80 1 de 1 al L i ~L AL adn 1 0 4 8 12 16 20 24 28 32 36 40 WEEKS FIGURE 5. Mean fasting serum-triglyceride levels for the five men with the high basal serum trigly- ceride and thirteen men with normal basal serum tri- glyceride concentrations. The dotted lines repre- sent the pre-dietary mean level for each group. From Roberts 1973. Reprinted with permission from Lancet. EFFECT OF DIETARY COMPOSITION ON DIABETIC COMPLICATIONS Other studies also have attempted to relate dietary patterns of a diabetic population with the types and frequency of complications occurring in them. In Rimoin's data on this subject (15), it was pointed out the clinical features of diabetes in various countries are difficult to correlate with dietary factors. Table 1 summarizes his findings. The frequency of juvenile- types of diabetes mellitus, as well as the general prevalence of vascular complications in chronic diabetes failed to incriminate relative dietary intakes of fat and carbohydrate in various ethnic groups. He suggested that it is not possible to separate dietary and other en- vironmental factors from genetic heterogeneity of the disease sufficiently to draw specific conclusions about diet and diabetic complications on a broad, epidemiologic basis. 302 Diabetes Mellitus TABLE 1. Ethnic Differences in Diabetes Mellitus* Diet Ketosis Vascular Fat Carbohydrate Complications European High High Common Common Rhodesian Sephardic Jew High High Uncommon Common Pima Indian High High Rare Common Alabama-Coushatta Indian High High Rare Common Seneca Indian High High Rare Common Navajo Indian High High Rare Uncommon Eskimo High Low Rare Rare Japanese Low High Rare Uncommon Ceylonese Low High Rare Uncommon Indian Low High Rare Common South African Indian Low High Rare Very Common South African Zulu Low High Common Rare Rhodesian African Low High Common Rare From Rimoin 1971. Reprinted with permission from Medical Clinics of North America. However, previous discussion in this monograph alludes to epidemiological surveys that take an opposing view about dietary influence on prevalence as well as complications of diabetes: Some relate sucrose intake to vascular disease; others the fat content of meal regimens. This controversy continues to be unresolved. EFFECTS OF DIET ON CONTRA-INSULIN HORMONES Theoretically, the ideal diabetic diet should promote efficient insulin action while avoid- ing stimulatory effects on other hormones known to oppose insulin. The latter include glucagon, another pancreatic islet hormone with potent hyperglycemic properties, adrenal cortisol, catecholamines, and pituitary growth hormone. It is known that the proportion of carbohydrate in the diet may influence day-to-day basal secretion of insulin (8) (Fig. 6). High carbohydrate diets maintain a higher ratio of insulin to glucagon plasma concentrations in association with optimum glucose tolerance, whereas low carbohydrate diets have the opposite effect on this hormonal ratio and are attended by reduced tolerance to glucose loads (19) (Fig. 7). These physicians have also reported that plasma glucagon is inappropriately elevated relative to glucose concentrations in the diabetic subject. Pure protein meals and certain amino acids stimulate secretion of insulin, glucagon, and growth hormone. Floyd and coworkers (5) have emphasized the synergistic effect between certain amino acids and glucagon on the pancreatic islet secretion of insulin, whereas Unger and co- workers (19) have reported inhibition of amino acid-induced glucagon release when glucose is administered concomitantly. In this same context, protein meals ingested before an oral glucose load improve glucose tolerance, but fat meals generally do not act in a synergistic fashion with glucose. High protein diets and very high sucrose-containing diets have been linked with ele- vated plasma levels of cortisol or increased adrenal cortisol secretory rates, which may explain the tendency toward impaired glucose tolerance reported in these situations. Manipulating portions, mixtures, and sequence of administration of nutrients does influence the balance between insulin and contra-insulin hormones and, perhaps, their summative action. Long-term effects of these dietary alterations, particularly with regard to the practical manage- ment of the diabetic patient, remain to be defined. Diet and Diabetes Mellitus 303 fo DIET 1 + DIET: —} Ad lib Low CHO High CHO Adlib LowCHO High CHO ow LG RR - - E EK I 4 A & A PLASMA 30 1 HY \ INSULIN I \ 3 ! PI AU/ml 1oF HE, { _ 1 WEIGHT,KG | ler A A L A Te A A Ad A A A A A A A RP [12 » ~ ¥ SOF I, ”° N dn 7 L { / Ne f 7 PLASMA 30Fg 0 — y INSULIN 5 “ ts Neo pU/ml 10} *v' S 80 1 wore Pes] ederatees SOHME . PLASMA 0 #* A FIGURE 6. Effect of high-carbohydrate and low-carbo- INSULIN / FN hydrate, isocaloric diets on basal plasma hyperinsulinemia pum a \ f of obese subjects. From Grey and Kipnis 1971. Reprinted a : pipe > with permission from the New England Journal of Medicine. WEIGHT, KG bh ale .o-vole-v ue 70 A A A A A A A A Er, oR we 5 \ / So A d \ vy . x X 1 PLASMA 3 A ~~ “o INSULIN vr? >] p/m ~ asf —! 105 WEIGHT, KG att lapel es A A 5 A A A A A A A A i. A bic or Be cB EB Ihe IE BLE BOE 8 WEEKS WRONG DIET OR POOR ADHERENCE TO DIET? The discussion up to this point has dealt with research aspects of diet and diabetes. How- ever, certain other major aspects of this subject are of more practical importance, and, to some degree, demand even greater attention. It is becoming more apparent than ever that another major deficiency of the dietary manage- ment of diabetic subjects relates directly to the failure of doctor and patient to ascertain whether or not the diet is actually being followed. In West's excellent review (20) of this problem, a number of publications were cited which indicate that adherence to diet by diabetic subjects is generally poor. In a British study only one-third of diabetic patients were found to consume within 10 percent of their prescribed cal- ories. Results of a national health survey in the United States in 1968 indicated that only 53 percent of diabetic subjects follow a physician's diet. The remainder were equally divided between those who were never given a meal plan or who simply had never adhered to one that was given to them. Even those who were ostensibly abiding by goals of good nutrition had a poor understanding of what sound dietary principles really are. 304 Diabetes Mellitus (MEAN t SEM.) GLUCOSE ! | 40 AU/ml : 20; INSULIN | o] : FIGURE 7. Fasting levels of glucose, insulin, and : glucagon during one week of normocaloric, low carbo- pg/ml : hydrate diet and during one week of a more balanced ISO GLUCAGON : normocaloric diet containing adequate quantities of ] carbohydrate. From Unger et al. 1971. Reprinted with permission from Transactions of the Association 100 of American Physicians. 509, - — ——— — > ccem—— ~ — —T CHO. 199 CHO. 193g PROT. 138 PROT. 86 od DIET par 189 DIET FAT 124 CAL 2313 CAL. 2260 r tox I 1 v T ¥ 1 T T 1 { 1 1 01 2 3456712 3465°6°T7 DAYS The failure of both physician and patient to achieve greater dietary precision in the con- trol of diabetes was very apparent in the data provided by the University Group Diabetes Pro- gram. Over 800 diabetic subjects were followed and evaluated in twelve different university- affiliated clinics for periods exceeding 8 years. The majority were obese. Despite carefully planned dietary regimens and the availability of expert clinicians, dieticians, and other medi- cal personnel, the vast majority of these subjects failed to register a significant weight loss and control of their diabetes at the conclusion of the program. This rather sad commentary on the poor success of optimum medical care facilities for con- trolling diabetes can be extended to nondiabetic obesity as well. At best, similar clinics that are specifically devoted to weight control have failed to correct obesity in the majority of cases when reporting is honest and follow-up periods span several months (17). The question is: Who is at fault, doctor, patient, or both? FUTURE DIRECTION OF RESEARCH The ideal diabetic diet, whether it is a single entity or a group of regimens fitted to various individual needs, has not yet been ascertained adequately enough by medical research. General epidemiological or retrospective studies of diet and diabetes to date prevent identifying the relative influence of genetic and environmental variables in sufficient detail to offer . definitive conclusions. This applies equally well to detailed clinical and laboratory research Diet and Diabetes Mellitus 305 carried out on a short-term basis, because the ultimate effect of specific regimens on control and the prevention of complications of the disease cannot be predicted on this basis. Such studies, nevertheless, do offer valuable clues and should continue with additional refinements and in conjunction with long-term investigations. One solution to this dilemma is a coordinated prospective study of diets of various types and their respective influence on carefully selected and monitored parameters. This requires a relatively large-scale, cooperative venture involving a number of patients and a number of medi- cal facilities. ORGANIZATION OF RESEARCH: PLANNING PHASE Experts in the fields of diabetes mellitus, nutrition, epidemiology, biostatistics, and ad- ministration are invited by the National Institutes of Health to plan the organization and types of research to be initiated in this program. Several approaches might be adopted, but basically they should include: 1) The establishment of a central administrative agency to coordinate the effort. 2) Procurement of a central data processing center for pooling of information and analysis of study results. 3) Development of a central laboratory to perform all critical laboratory examinations or else provide guidelines and quality control for procedures performed in other centers. 4) Selection of participating medical centers throughout the country whose existing re- sources, facilities, and manpower would allow an in-depth investigation of these research prob- lems. Selection might be based on competitive research, center, or contract grant systems. 5) Definition of patient selection, types of research, parameters to be monitored, etc. after appropriate exchanges with principal representatives of participating medical centers dur- in the planning phase. ORGANIZATION OF RESEARCH: RESEARCH PHASE Patients participating in this study are selected on a voluntary basis with each individual being made fully aware of the nature and purpose of the study and acknowledging all aspects of it through written, informed consent. Selection of participants is one of the most critical factors in the study. It is the author's opinion that diabetic patients should be free of major complications of the disease and in relatively good health as determined by initial history, physical examinations, and laboratory screening procedures. A portion of subjects in any one of the study groups serves as a control population and is given a standard diabetic diet consisting of 40 percent carbohydrate, 40 percent fat, and 20 percent protein and of sufficient calories to maintain ideal body weight. Obese diabetic sub- jects are reduced to ideal body weight before being considered for this study. Extreme care is exercised to insure maintenance of ideal weight throughout the investigation. Other groups of patients are given diets to test the effects of varying distribution, com- position, and sequencing of nutrients on a variety of pertinent metabolic profiles. For example, some studies would focus on diets relatively high in carbohydrate content; others might examine the outcome of low fat intake or high protein consumption. Types of nutrients also must be assessed such as regimens comparing different kinds of carbohydrate (e.g., starch, fruit 306 Diabetes Mellitus carbohydrate, and more refined sugars) or different types of fat (polyunsaturates vs saturated dietary lipid). Finally, the influence of timing and combinations of nutrients should be evalu- ated. One might examine the number or frequency of meals and the order of administration of specific foods. The latter would include comparisons of mixed meals with isolated intake of specific nutrients separately and in various sequences. Patients report to designated outpatient centers at regular intervals. During this time ap- propriate entries into standardized records are performed and with reference to physical findings and laboratory data. Nutritional and medical counseling is an integral part of these visitations. Parameters to be monitored in the laboratory include diabetic control as evidenced by basal blood glucose, cholesterol, and other lipid levels. These, in turn, might be correlated with plasma concentrations of other fuels including amino acids, ketone bodies, free fatty acids, etc. The influence of diets on hormonal profiles is also ascertained. Insulin dynamics are related to concentrations of circulating contra-insulin hormones such as growth hormone, epinephrine, glu- cagon, cortisol, etc. At various intervals patients are hospitalized for very brief periods to determine how various dietary regimens influence these parameters throughout a 24-hour period, both during and between meals. One could envision initial prospective studies proceeding on a relatively short-term basis for a 6- to 12-month period. Pooled data analyses might dictate which regimens show greatest promise for optimum control as compared to more standard regimens in use today. Such regimens could be adopted for long-term studies (5-10 years) while continuing standard diets in control groups as a basis for reference. In long-term studies onemight scrutinize in greater detail physical evidence for development of diabetic complications, such as microvascular diseases of the eye and kidney, peripheral and cardiovascular disease, and neurologic changes, which are parameters that are more likely to change over more extended periods of time. These, in turn, are related statistically to labora- tory data derived from analyses of blood tests. The value of this research has relevance to health beyond that of diabetic patients. Nutri- tional information such as this very likely would apply to nondiabetic populations as well, since the high incidence of vascular disease in this country undoubtedly reflects dietary habits of citizens without diabetes as well. In this regard, long-term studies might very well involve ad- ditional control groups who do not have diabetes mellitus, but who might also be examined in a systematic way to evaluate diet and its effect on ultimate development of cerebral, cardiac and peripheral vascular disease, and other indices of morbidity and mortality. Information derived from short-term studies could conceivably r&luce the costs of similar long-term studies by as much as 50 percent annually, since many types of regimens could be elim- inated and attention placed on dietary approaches of fewer number. The final outcome of this research is the development of meal regimens that promote optimum control of blood glucose and lipids, minimal aggravation of metabolic and hormonal factors that oppose insulin action, and maximum suppression of events leading to vascular, renal, and neuro- logical complications of diabetes mellitus. FUTURE GOALS OF MEDICAL EDUCATION IN THE FIELD OF NUTRITION It is unfortunate that very few physicians have a great understanding of nutrition generally and the dietary aspects of diabetes and obesity management. It is appropriate, then, to care- fully examine present day curricula in our medical schools, other paramedical professional Diet and Diabetes Mellitus 307 schools, and postgraduate training programs in order to assess the seriousness of the deficiency. In this regard, groups whose charge is improvement of medical education might take specific steps to evaluate the situation. In this same context, pilot programs in nutritional education might be ‘initiated in selected universities and subsidized by the Federal government. The costs incurred would very likely ap- proach the cost of supporting a small department or section within a medical school for person- nel, teaching aids, and equipment. The adequacy of these programs could be evaluated directly by comparing performances on board examinations of medical and other professional students who have been exposed to this type of education to board scores of those who have not been exposed. The value of this type of ap- proach would be to improve the expertise of future practicing physicians in the nutritional management of patients with diabetes and related disorders, as well as in other forms of disease requiring similar dietary therapy. FUTURE GOALS OF DIABETIC CARE FACILITIES The education of patients with diabetes and/or obesity about nutrition is a team effort. In its broadest scope, it is a multi-institutional objective involving physicians, dietitians, nurses, hospitals, lay groups, news media, industry, labor unions, and public health agencies. At this point in time, however, dissemination of information about sound nutrition is often fragmented and duplicated, and its impact is not felt by the general public or else it is ob- sucred by more sensational types of food fads that receive greater publicity, despite their lack of acceptance by the medical profession. The American Diabetes Association and TOPS Club, Inc. (Take Off Pounds Sensibly) are two nonprofit organizations devoted to the care and management of individuals with diabetes and obesity, respectively. Their combined membership totals several hundred thousand individuals. Their success and those of other self-help groups are already well known and often compare favor- ably with results of more expensive, less practical specialty clinics (17). They look to the medical profession for guidance and have already taken steps to improve communication along these lines. Their own publications reach virtually every state in the union and several countries throughout the world at the local chapter level. Public health officials concerned about nutrition for the diabetic or the obese individual should take advantage of existing resources of these self-help groups. One major step forward would be to develop a more extensive flow of information derived from medical research and opinion or from guidelines and viewpoints expressed by public health agencies to these nonprofit organizations. In this regard an interlocking council could coordinate such an effort. The council might be composed of representatives from professions and lay societies and public health organizations whose specific responsibility is informing the public in the most efficient manner about nutritional principles for the control of diabetes and obesity. The Federal government should also encourage the development of more comprehensive informa- tion centers within nonprofit lay societies who are already helping the diabetic and obese sub- ject. Subsidization of activities concerned with release of information about medically accepted nutritional principles at meetings, through publications, other news media; the establishment of telephone '"hotlines' for physician and patient needs; the institution of reference libraries to aid physicians, paramedical personnel, and the general public in their quest for specific nutri- 308 Diabetes Mellitus tional information are all means of bringing the public close to the facts at the lowest possible cost. Ultimately, these activities may insure the development of more organized, well-informed, and effective programs in communities for purposes of weight control and diabetes management. Such preventive medicine will benefit public health to a great degree, since the number of in- dividuals with either obesity, diabetes, or both has reached significant proportions and consti- tutes a major health hazard in this country today. REFERENCES 1. Bierman, EL, and D Porte, Jr 1968. Carbohydrate intolerance and lipemia. Ann Int Med 68:926-933. 2. Bierman, EL, MJ Albrink, RA Arky, WE Connor, S Dayton, N Spritz, and D Steinberg 1971. Principles of nutrition and dietary recommendations for patients with diabetes mellitus: 1971. Diabetes 20:633-634. 3. Brunzell, JD, RL Lerner, WR Hazzard, D Porte, Jr, and EL Bierman 1971. Improved glucose tolerance with high carbohydrate feeding in mild diabetes. New England Journal Med 284: 521-524, 4. Campbell, GB 1963. Diabetes in Asians and Africans in and around Durban. South Afr Med J 37:1195-1208. 5. Floyd, JC, Jr, SS Fajans, S Pek, CA Tiffault, RF Knopf, and JW Conn 1970. Synergistic effect of essential amino acids and glucose upon insulin secretion in man. Diabetes 19:109-115. 6. Fredrickson, DS, and RI Levy 1972. Familial hyperlipoproteinemia. In Metabolic basis of inherited disease (Third edition). Eds, JB Stanbury, JB Wyngaarden, and DS Fredrickson. New York, McGraw-Hill, p 602. 7. Goto, Y, Y Nakayama, and T Yagi 1958. Influence of the World War II food shortage on the incidence of diabetes mellitus in Japan. Diabetes 7:133-135. 8. Grey, N, and DM Kipnis 1971. Effect of diet composition on the hyperinsulinemia of obesity. N England J Med 285:827-831. 9. Himsworth, HP 1949. The syndrome of diabetes and its causes. Lancet i:465-472. 10. Kalkhoff, RK, HJ Kim, J Cerletty, and CA Ferrou 1971. Metabolic effects of weight loss in obese subjects. Changes in plasma substrate levels, insulin growth hormone responses. Diabetes 20:83-91. 11. Kaufmann, NA, and Y Stein 1970. Carbohydrate induced hyperlipemia. Atherosclerosis 11: 365-367. 12. Meinert, CL, GL Knatterud, TE Prout, and CR Klimt 1970. A study of the effects of hypoglycemic agents on vascular complications in patients with adult-onset diabetes II. Mortality results. Diabetes 19 (Suppl 2):789-830. 13. Poon-King, T, MV Henry, and F Rampersad 1968. The prevalence and natural history of dia- betes in Trinidad.. Lancet i:155-160. 14. Reid, JM, SD Sullmer, KD Pettigrew, TA Burch, PH Bennett, M Miller, and GD Whedon 1971. Nutrient intake of Pima Indian women: Relationships to diabetes mellitus and gallbladder disease. Am J Clin Nutr 24:1281-1289. 15. Rimoin, DL 1971. Inheritance in diabetes mellitus. Med Clin North Am 55:807-819. 16. 17. 18. 19. 20. 21. Diet and Diabetes Mellitus 309 Roberts, AM 1973. Effects of a sucrose-free diet on the serum-lipid levels of men in Antarctica. Lancet i:1201-1204. Stunkard, A, H Levine, and S Fox 1970. The management of obesity. Patient self-help and medical treatment. Arch Int Med 125:1067-1072. Sukhatme, PV 1969. On the trend of obesity in advanced countries. In Diabetes. Proceed- ings of the Sixth Congress of the International Diabetes Federation. Ed J Ostman, and RDG Milner. Amsterdam. Excerpta Medica Foundation Press, pp 704-714. Unger, RH, WA Muller, and GR Faloona 1971. Insulin/glucagon ratio. Trans Ass Amer Physcns 84:122-129. West, KM, and JM Kalbfleisch 1971. Influence of nutritional factors on diabetes prevalence. Diabetes 20:99-108. Yudkin, J 1969. Dietary fat and dietary sugar in relation to ischemic heart disease and diabetes. Lancet ii:4-5. 23 INSULIN SYNTHESIS AND ANALOGS Harold E. Lebovitz INTRODUCTION The major defect in diabetes mellitus is a lack of normal metabolic regulation by insulin. Insulin controls many vital cellular processes and acts either in concert with, or in opposition to, several other regulatory hormones (glucagon, growth hormone, catecholamines, and glucocorticoids). Research during the last two decades has provided significant insight into an understanding of insulin secretion and action. Utilizing this information, it is reasonable to develop a schema that envisions diabetes mellitus as not being caused by a single defect but resulting from any one of a variety of biochemical disturbances. Figure 1 illustrates at least eight or nine possible metabolic abnormalities that could result in the disturbance of carbohydrate and lipid metabolism that we call diabetes mellitus. Defect 3, which is an absolute absence of insulin synthesis in the beta cell, is the abnormality noted in the juvenile form of diabetes mellitus (63). Defects 1 and 2, either or both of which would result in an inadequate amount of insulin being released even though significant pancreatic stores are present, is the type of abnormality seen in patients with adult onset diabetes mellitus (28,31,64). Defect 4, an alternation of hepatic action of insulin, would be an inability of insulin to properly inhibit the enzymes of gluconeogenesis and stimulate those of glycolysis and glycogenesis, resulting in unabated hepatic glucose production even in the presence of insulin (41). Such a defect may account in part for the uncontrolled hyperglycemia seen in nonketotic hyperosmolar coma. Defect 5 is an alteration of the binding, or rate of destruction of insulin, in either passage through the liver or circulating in the blood. Defects 6, 7, and 8 are lack of insulin action because of either deficient insulin receptors, deficient production of an intracellular messenger, or inability of the messenger to affect the final biochemical intra- cellular effectors. While no definitive proof of these types of defects (6 through 8) exists, there are several animal (sand rat) (18) and human (lipoatrophic diabetes mellitus) (43) types of diabetes mellitus that could be explained by such mechanisms. Defect 9, caused by insulin antagonistic hor- mones, is probably mediated through mechanisms 6 to 8. Since all forms of diabetes mellitus are ultimately associated with lack of insulin action, therapeutic endeavors naturally center upon developing agents which either exert insulin action or increase the release of endogenous insulin. It must be borne in mind, however, that other therapeu- tic agents which either facilitate endogenous insulin action or stimulate one or more of the bio- chemical effects of insulin, may be equally useful. The present section is devoted to an analysis of the potential benefits of synthetic insulin peptides and chemically modified insulins in the treatment of diabetes mellitus and other metabolic disorders. In order to do this, it is necessary to first review the chemical nature of insulin, the spectrum of insulin's biologic activities, the relationship of insulin to other anabolic agents or growth factors, and the mechanism by which insulin exerts its actions. Since the availability of synthetic and modified insulins depends on peptide synthetic methodology, some attention must necessarily be focused on this area. 310 Insulin Synthesis and Analogs 311 ® INSULIN ANTAGONISTIC HORMONES 0 ® OR M— Om ® cIreuLATING Lr? INSULIN CELL © STIMULATORY AGENTS (substrates, hormones, ions) ® INSULIN — PANCREAS LIVER © INHIBITORY AGENTS (Monoamines) FIGURE 1. Biochemical defects that could cause the carbohydrate and lipid abnormalities associated with diabetes mellitus. Defects 1, 2, and 3 are associated with decreased re- lease of insulin from the pancreas. In defect 1, insulin secretion is impaired because the stimulatory release mechanism is disturbed. In defect 2, release is impaired because of the presence of agents that inhibit the secretory process. Defect 3 is absent or abnormal insulin synthesis so that the pancreas does not contain any normal insulin to be secreted. Defect 4 is failure of the liver to respond to insulin. Defect 5 is excessive destruction of insulin either in passage through the liver or circulatory system. Defect 6 is an abnorm- ality of the insulin receptor so that insulin does not bind. Defect 7 is failure to generate a secondary messenger after insulin attaches to the receptor. Defect 8 is failure of the gen- erated intracellular messenger to effect the appropriate response inside the cell. Defect 9 is antagonism of insulin action by other hormones and is probably due to events caused at sites 6 to 8 by those hormones. INSULIN STRUCTURE Since insulin is a small, readily available, and biologically important protein, it has served as a model for the development of much of the methodology of modern protein chemistry. Insulin was the first protein to have its amino acid sequence completely determined (Ryle et al., 1955). The development of methods for determining N-terminal amino acids indicated that insulin consisted of two chains: one with an N-terminal glycine and one with an N-terminal phenylalanine (48). It con- tains three disulfide bridges which contribute to its three-dimensional structure. Table 1 depicts the primary amino acid sequence of beef insulin. The primary sequence for more than 20 vertebrate insulins have been completely determined, and the data indicate that 21 of the 51 amino acid residues (including the 6 cysteines) are invariant. Table 3, taken from Blundell et al. (3), summarizes these data. A major achievement in opening new vistas into understanding the chemical basis of insulin action has been the elucidation of its three-dimensional structure (Fig. 2). X-ray crystallography has allowed delineation of the structure of the insulin monomer, diamer, and hexamer. Complete descrip- tions are available in several recent reviews (2,3,21). The most obvious stabilizing forces result from the disulfide bonds. The A7-B7 bond is on the outside of the molecule, while B19-A20 is more concealed but still somewhat accessible to the solvent. The A6-All disulfide bond is completely buried within the molecule and forms part of the nonpolar core. A second important feature of the monomer structure is the existence of a completely nonpolar core. This hydrophobic center consists TABLE 1 AMINO ACID SEQUENCE OF BOVINE INSULIN A CHAIN 5 Gly-Ile-Val-Glu-Gln-Cys-Cys-Ala-Ser-Val-Cys-Ser-Leu-Tyr-Gln-Leu-Glu-Asn-Tyr-Cys-Asn 1-2-3 4 5-6 7 § 9.10 112 22:13 4 15 16 17 18 19 20 2) S / B CHAIN Phe-Val-Asn-Gln-His-Leu-Cys-Gly-Ser-His-Leu-Val-Glu-Ala-Leu-Tyr-Leu-Val-Cys-Gly-Glu-Arg-Gly-Phe-Phe-Tyr-Thr-Pro-Lys-Ala 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 0 SN3ITTTON S939qRTa ZIE TABLE 2 THE KNOWN AMINO ACID SEQUENCES OF INSULIN A CHAIN 1 2 3 4 5 6 7 8 5 10 11 12 13 14 15 16 17 18 19 20 21 Mammals ©1Y Ile Val Glu Gln Cys Cys Thr Ser Ile Cys Ser Leu Tyr Gln Leu Gln Asn Tyr Cys Asn Asp Ala Gly Thr Val Birds His Asn Thr Fish Leu His His Pro Asn Lys Phe Asp Gln Ser Lys Asp Ile Asn Arg Cuings Pag Asp Ala Gln Thr Thr Arg His Gln ypu Asp Thr Asn Ile Arg Asn Met Asp B CHAIN 0 1 2 2 4 5.6 7 8 9 10°11 12 13 14 15 16 17 18 19 20 21 22 22 24 25 26 27 28 2 230 Phe Val Asn Gln His Leu Cys Gly Ser His Leu Val Glu Ala Leu Tyr Leu Val Cys Gly Glu Arg Gly Phe Phe Tyr Thr Pro Lys Ala Pro Mammals Lys Met Ser Thr Birds Ala Ala Fish Met Ala Ala Ala Asp Asp Asn Ser Lys - Ala Pro Pro = Val Guinea pig Ser Arg Asn Thr Ser Gln Asp Asp Ile Lys Asp Coypu Tyr Ser Arg Gln Asp Thr Ser Arg His Arg Tyr Arg Pro Asn Asp - £T¢ sborTeuy pup STS8Y3uAg UTTNSUT 314 Diabetes Mellitus of the residues B6, B11, B15, and Al6 leucines; B18 valine; B24 phenylalanine, and the phenyl ring of B26 tyrosine; the A6-All and part of the A20-B19 cysteines. A third set of stabilizing forces helping to maintain the three-dimensional structure are ion pairs and hydrogen bonds. Evidence suggests that the C-terminal carboxyl group of A21 and the guanidinium group of R22 arginine, and the B29 lysine a amino group and A4 glutamate carboxyl group form ion pairs. The complete insulin molecule is very compact and only the carboxyl and amino terminals of the B chain extend out from the main structure. The surface of the molecule consists of two small nonpolar regions which are involved in diamer formation, but the major portion of the surface consists of polar (hydrophilic) residues. Zinc binding occurs through the B10 histidine and leads to the formation of hexameric crystals. FIGURE 2. Structure of insulin monomer (from Blundell et al. 1971). Reprinted with premission from Recent Progress in B2l Hormone Research. From the three-dimensional model, it is of interest to note that t:he invariant amino acid residues of naturally occurring insulins (as noted from Table 2) may be grouped into three cate- gories: (1) Those residues responsible for the backbone of the monomer structure, i.e., all three cystine groups--glycine B8 and B23, leucine B5, Bll, B15, and Al5, valine B18 and isoleucine A2; (2) those polar and nonpolar residues involved in diamer formation--.serine B9, valine B12, tyrosine B16, and phenylalanine B24; and, (3) certain polar residues lying near one another on the surface of the A chain--glycine Al, glutamate A5, tyrosine Al9, and asparagine A21. Insulin Synthesis and Analogs 315 STRUCTURE-ACTIVITY RELATIONSHIP OF THE INSULIN MOLECULE The chemical approaches to defining the structure-activity relationship of the insulin mole- cule have involved chemical modifications, selective degradation by enzymes or chemical reagents, peptide synthesis, or a combination of these procedures. For structure-activity studies to be meaningful, several criteria must be met. The insulin derivative must be isolated and the alter- ation specifically characterized. The effect of the alteration on the three-dimensional structure of the insulin molecule must be determined. The most readily available procedures for determining whether the three-dimensional structure of the molecule has been altered is to measure whether the insulin derivative rotates polarized light in a pattern different from that of the native molecule. Two techniques which have been useful in measuring the optical properties of the insulin molecule and its derivatives are optical rotary dispersion and circular dichroism. Finally, the methods for measuring biological activity must be sensitive and precise and should assess the entire spectrum of biological activities. Early studies on insulin structure-activity relationships involved chemical modifications in which the specific products were not characterized. These studies showed that limited acylation of the a and e amino groups, esterification of the side chain hydroxyls, and, to some extent, blocking of guanido and imidazolyl groups have little or no effect on the measured biological activity. In contrast, extensive esterification of the free carboxyl groups, diazotization of the histidine and tyrosine residues, excessive iodination and reductive splitting of the disulfide bonds causes definitive loss of biological activity. Complete reviews of the studies of chemical modifications on insulin activity have recently been published (3, 23). Structure-activity studies in the last 10 years have utilized the basic knowledge derived from the amino acid sequence and three-dimensional structure data and have relied heavily on very specific chemical modifications, selective enzyme degradation, resynthesis of insulin-like com- ounds from selectively degraded insulins, synthesis of insulin chains or analogs, and combination of these natural chains or synthesized chains (68). From the results of all these studies, several major concepts concerning the structural basis of insulin action can be made. Those modifications which cause the greatest change in three-dimensional structure (circular dichroism or optical rotary dispersion changes) are associated with the greatest loss of activity (3). It is not possible to define a specific active site of the insulin molecule (3, 23). Specific aspects of structure-activity relationships that appear to be important are: A. Disulfide Groups: Disruption of all three disulfide bonds leads to loss of all activity (26, 66). The least protected disulfide group (A7-B7) can be specifically reduced and either carboxymethylated or S-sulfonated (8). As noted in Tgble 3, this partially reduced insulin has significant biological activity. Weitzel and co-workers have synthesized an insulin analog in which cysteine A7 is replaced by alanine, and it is still active (23). The most protected disul- fide bond A6-All is buried in the nonpolar core of the molecule. If both cysteines are replaced by alanine, the resultant synthetic analog is inactive (55, 56). If, however, the disulfide bridge is replaced by a trioether bridge (A11-A6 cystathione insulin), insulin activity is not lost and circular dichroism is unchanged, indicating little or no change in the conformation of the molecule (24, 25). 316 Diabetes Mellitus TABLE 3. Insulin Analogs with Alterations of Disulfide Groups! Percent Biological Activity Percent Fat cell Mouse Convulsion Rat Hypoglycemia Immunoreactivity [b, (S-carboxymethyl cystein) ,7 7] 40 100 [Di (S-sulfocysteine),; py] 15 4 to 10 5.6 to 11,3 Carba IN Insulin (A 11-A6 cystathione insulin) "Active" "Active" 1Zahn et al. 1972. Diabetes 21 (suppl 2) 468-475. Reprinted with permission from Diabetes, the Journal of the American Diabetes Assn. B. Alpha and Epsilon Amino Groups: Many studies have been carried out with reagents inter- acting with the free amino groups of the molecule, and the results are somewhat dependent on the nature of the reacting group. Table 4 summarizes results obtained with a few derivatives (see (3) for complete list). Alterations in Bl phenylalanine do not alter the three-dimensional structure of the molecule, however, deletion or alteration of Al glycine does change circular dichroism significantly. Alteration of the Al glycine leads to a remarkable loss in activity, whereas loss or alteration of Bl phenylalanine has little effect on activity (4, 68). Elongating the A chain significantly decreases activity (68). Additional synthetic studies by Weitzel and his group have shown that the sequence Al-3 Gly, Ile, Val is necessary for activity as replacement of Gly Al by Ala or BAla, of Ile A2 by Ala or Leu, and of Val A3 by Leu or Ile, lead to an almost complete loss of activity (22, 53, 56). CARBOXYLIC ACID SIDE CHAINS AND TERMINAL RESIDUES The six carboxylate groups in insulin lie on the surface of the monomer. The A4 glutamate and A21 carboxylate probably form salt bridges with B29 lysine and B22 arginine, respectively (3). Total esterification of all the carboxyl groups results in conformational change in the molecule and loss of biological activity (32). Table 5 lists the data that indicate that a carboxyl at A21 is very important in maintaining structure and activity. The B terminal alanine and several adja- cent residues play little or no role in structure or activity, but the region B23 to B27 is very important for molecular activity. B22 ARGININE SIDE CHAIN With the exception of guinea pig insulin, all insulins have an arginine at B22. Weitzel et al.’ (59) synthesized B chain analogs using the Merrifield solid state synthesis technique and recom- bined them with natural A chain. Replacement of B9 and B27 by alanine yielded an analog with 75 percent activity as assessed by the mouse convulsion assay. When insulin analogs were made which had alanine at B9 and B27 and either histidine or alanine at B22, their activities were less than 0.5 percent that of native insulin. On the other hand, replacement of arginine B12 by ornithine gives a derivative with 16 percent of the activity of the synthetic analog with arginine at B22. These results are consistent with the presence of a B22 guanidinium group, or other positive ion, salt bridge with A21 carboxylate, which is important in maintaining the structure of the molecule. Unfortunately, studies of the conformation of the B22 modified insulins are not available. Insulin Synthesis and Analogs 317 TABLE 4. Insulin Derivatives and Analogs with Alterations in the N Terminus and B29 eAmino Lysine! Percent Biological Activity Percent Fat Cell Mouse Convulsion Rat Hypoglycemia Immunoreactivity Des Amino Al 15 - 40 Des Gly Al 2t010 10 20 to 30 Des Phe Bl 90 103 Acetyl Al 40 100 Acetyl B29 75 100 Diacetyl Bl, B29 85 100 Arg Gly Al y 68 59 Lys Arg Gly Al 20 [Gly Bl] eacetyl B29 35 25 Des Gly Al des Phe Bl 1.6 7 15 Des (Gly Al Ileu A2) 0 0 Des (Phe Bl Val B2) 0.2 2 ‘17ahn et al, 1972, Diabetes 21 (suppl 2) 468-475. Reprinted with permission from Diabetes, the Journal of the American Diabetes Assn. TABLE 5. Insulin Derivatives with Modifications at the Carboxyl Terminus Percent Percent Conformational Biological Activity Immunoreactivity Change Fat Pad Mouse Convulsion = Desoctapeptide Insulin! 0.7 <1 3 Marked (B23 to B30 deleted) Des Ala B302 100 None Des Ala B30 des Asn A212 5.5 4 5 Marked Methylated Insulin? <3.0 Marked Des Amido A212 - 100 [Glu 5, Ala Al2, Phe Al9, Ala A21]% Insulin 100 Des Ala B30 des Lys B29 des ‘Pro B28]5 Insulin 84 lRager et al. 1969 (44) 2Carpenter 1966 (10) 3Levy and Carpenter 1970 (32) “Weitzel et al. 1968 (56) SKatsoyannis 1969 (26) 318 Diabetes Mellitus TABLE 6. Actions of Insulin Nucleic Acid Metabolism 1 2 3. 4 5 Increases uridine and cytidine transport or phosphorylation in bone cells (42) Increases DNA synthesis (49) Promotes cell divistion (49) Facilitates differentiation of muscle cells (13) Increases synthesis of t-RNA (36) Protein Metabolism As 2. 3. 4. Increases amino acid transport (45) Increases amino acid incorporation in protein (37,60,62) Increases protein synthesis by ribosomes (61) Decreases protein catabolism (38) Glucose Metabolism 1. 2+ 3. 4. Increases glucose uptake by muscle, adipose tissue, and liver (45) Increases UDPG-o glucan transglucosylase activity (30) Decreases gluconeogenesis (35) Increases glycolysis Lipid Metabolism 1. 2. 2. 4. Increases synthesis of free fatty acids (33) Decreases lipolysis (15) Decreases ketone formation (17) Increases esterification of free fatty acids (1) Increases Transport of Potassium into Cells Decreases Cyclic AMP in Some Tissues (9) Insulin Synthesis and Analogs 319 OTHER SIDE CHAINS AND AMINO ACID RESIDUES Many studies of the role of the tyrosine residues on the biological activity of insulin have been done but the data, for the most part, are conflicting. The Al4 and Al9 tyrosine residues are on the outside of the monomer, diamer, and hexamer. The B26 tyrosine is on the surface of the monomer, but is buried on the inside of the diamer and hexamer. The B16 tyrosine is fully exposed in the monomer, but partially buried in the diamer (3). The properties of several of the tyrosines are anomalous, probably due to their specific environments. For nitration reactions at pH 7 to pH 3, the reactivity is: Tyr Al4 > Tyr Al19 > Tyr B16 > Tyr B26. The mononitro (Al4 or A19) and dinitro (A19 and Al4) insulins have 104 + 10 percent and 74 + 6 percent activity, as compared to insulin as measured in blood glucose depression in chronically diabetic mice, but only 53 + 10.8 percent and 25 percent in the mouse confulsion test (39). Tetranitrotyrosine insulin is reported to have 50 percent activity in the mouse convulsion test (3). These data, which indicate that the tyrosines are probably not necessary for activity, have been confirmed by the synthesis of insulin analogs in which some tyrosine residues have been replaced and biological activity maintained [Tyr Al4 can be replaced by Phe or Ala (55, 56); Tyr Al9 can be replaced by Phe but not Ala (56)]. Iodination of the insulin molecule gives a variety of products, many of which have markedly reduced activity. While the B10 histidine does not appear to be essential for insulin activity, there is some data which indicate that His B5 may be important for biologi- cal activity (58) but additional studies are necessary to substantiate this thesis. The aliphatic hydroxyl groups Thr A8, Ser A9, Ser Al2, Ser B9, and Thr B27 are on the surface of the diamer. Semisynthetic analogs [Glu A5, Ala Al2, Ala Al8, Ala A23] insulin (56) and [Ala B9, Ala B27, desAla B30, desLys B29, desPro B28] insulins (57) are reported to have 75 percent the activity of native insulin. These data indicate that Ser Al2, Ser B9, and Thr B27 are not impor- tant for biological activity. BIOLOGICAL ACTIONS OF INSULIN Since the early 1950's, there has been an extensive inquiry into the nature and mechanism of insulin actions. These studies have led to the realization that insulin exerts numerous actions at many different biochemical loci. The net effect of all these actions is to facilitate the storage and utilization of substrates and to promote growth and differentiation. Table 6 lists the actions of insulin. This list is not exhaustive, but does include most of the actions that have been reasonably well studied. Many of these actions of insulin are interrelated. For instance, insulin facilitates the entry of glucose into muscle and liver cells where it is either metabolized via glycolysis, stored as glycogen, or converted to free fatty acids. As insulin acts on the liver to increase glucose uptake and metabolism, it also blocks gluconeogenesis and ketogenesis. The insulin facilitated entry of glucose into adipose tissue cells causes increased fatty acid synthesis and esterification of fatty acids to triglycerides. Likewise, many of the anabolic actions are interrelated. There is much evidence to indicate that the different actions of insulin are not mediated through the same mechanism. Some of its actions involve transport across the cell membrane (glucose uptake, amino acid transport, uridine and cytidine uptake), while others involve alter- ations in intracellular processes (increased DNA and RNA synthesis, increased protein synthesis by ribosomes, enzyme induction). Some insulin actions are thought to be mediated by a decrease in 320 Diabetes Mellitus tissue cyclic AMP (effects on isolated rat liver; antilipolytic effects) but this is not unequi- vocally proved, and there are many actions for which this mechanism is not likely (glucose trans- port, enzyme induction, protein synthesis). Insulin has been shown to alter the activity of a number of enzymes (29). At least two mechanisms have been described by which insulin alters enzyme activity. One involves the dephosphorylation of a protein enzyme (glycogen transferase, adipose tissue lipase). The other involves de novo synthesis of the enzyme (glucokinase, hexokinase II of adipose tissue, tyrosine aminotransferase of hepatoma cells in tissue culture). Insulin stimula- tion of protein synthesis can occur either by stimulation of transcription (glucokinase, hexo- kinase II) or translation (tyrosine aminotransferase of heptoma cells in culture). RELATIONSHIP OF INSULIN TO OTHER ANABOLIC HORMONES Insulin is essential for anabolism to occur in mammalian systems. As noted in the preceding section, insulin stimulates many processes that are involved in tissue growth and development. There have been described in the last few years a number of other anabolic peptide hormones which seem to be related to insulin. Growth hormone, which is secreted from the anterior pituitary gland and seems to play the major role in controlling linear growth, has been shown to exert some, if not perhaps all, of its anabolic activity through a second circulating protein whose synthesis it seems to control (12). This second hormone has been named sulfation factor, or more recently, somatomedin. Somatomedin is responsible for the stimulation of cartilage macromolecular synthesis and growth. Somatomedin has not been completely purified and characterized. Partially purified preparations, however, have some insulin- like activity on adipose tissue [antilipolytic (51) and increase l“C, glucose oxidation to l%CO0, (19)], and according to Hintz et al. (20) interact with the insulin receptors of adipose tissue, liver cells, and chondrocytes. Of additional interest are the observations that very high concen- trations of insulin will mimic the effects of somatomedin on cartilage macromolecular synthesis in vitro (47). These data suggest that insulin and somatomedin may be related. The growth of autonomic nerves has been shown to be stimulated by a factor which is present in the salivary glands of male mice. This factor, which is called nerve growth factor, has been puri- fied and its entire amino acid sequence determined (16). Bradshaw and co-workers have suggested that nerve growth factor is structurally related to proinsulin (16). Nerve growth factor has been shown to facilitate growth of autonomic ganglion and sensory nerves in embryos and in tissue culture. It has not been shown to have insulin-like actions, but insulin in high concentrations can stimulate the growth of neurons sensitive to nerve growth factor. Insulin and nerve growth factor have simi- lar effects in their respective target tissues (16). Another growth factor which is found in the submaxillary glands of mice is epidermal growth factor (50). This peptide has been isolated and structured. It facilitates anabolic reactions in tissues derived from epidermis. The relationship of insulin to these other tissue growth hormones promises to be intriguing and may open up new areas of information which will allow us to understand the control of growth in specific tissues. INSULIN RECEPTORS Insulin action on adipose tissue, liver, and possibly other cells, appears to result from the interaction of insulin with a specific receptor binding site on the external surface of the cell Insulin Synthesis and Analogs 321 membrane (11). The current theories presume that after insulin binds to the receptor, it may alter the conformation of the membrane, thereby allowing changes in hexose, amino acid, and possi- bly ion transport. Other consequences of the binding of insulin to the receptor may be the release of one or more second messengers that are released and mediate the intracellular effects of insulin. The decrease of cyclic AMP that is the consequence of the action of insulin on some cells could be due either to some inhibition of adenyl cyclase by either a change in the membrane or a messenger which blocks the cyclase, or possibly, the activation of an intracellular phosphodiesterase by some second messenger. Many studies have confirmed the existence of the insulin receptor and have demonstrated that insulin exerts its actions even though it cannot penetrate into the cell. The events subsequent to insulin-receptor interaction are still speculative, with little definitive proof. DIRECTIONS FOR FUTURE RESEARCH Future research needs in the area of the therapy of diabetes mellitus with synthetic insulin and insulin analogs must be subdivided into immediate and long-range goals. A. Immediate Needs: The major immediate need is the development of more efficient, more rapid, and cheaper methods for the synthesis of insulin and insulin analogs. Since the A chain is 21 residues and the B chain is 30 residues, the synthesis of the chains presents major problems in obtaining high yields of purified chains. Classical techniques of organic peptide synthesis are very laborious and require extensive purification of each subunit of the chain before it can be coupled with the next subunit. The final yield of chains from the starting material represents only a few percent. The Merrifield solid state synthesis method has been used by several investi- gators, but the products are heterogenous and very difficult to purify. An immediate need is in- vestment in basic research to develop newer and better techniques for synthesizing peptides. This research will clearly be of benefit in the therapy of many diseases other than diabetes mellitus, since synthetic peptide analogs of many hormones have now been shown to have therapeutic benefits. Another major problem in the synthesis of insulin and its analogs is the combination of the A and B chains to give high yields of purified synthetic insulin. The techniques worked out by Zahn and co-workers (65), Katsoyannis and Tometsko (27) and the Chinese group [Du et al. (14)] give yields of 2 to 10 percent insulin from combination of the chains (34). This low yield and the contamination with all sorts of erroneous chain combinations make the problem of the synthesis of large quantities of insulin almost insolvable. As a result of knowledge obtained from the three-dimensional structure of insulin, which indicates that Gly Al and Lys B29 are separated by 8 to 10 angstroms, and previous structure-function studies that showed that the presence of acetyl groups at Gly Al and Lys B29 did not affect the hypoglycemic activity of insulin, Brandenburg (4) prepared a number of modified insulins in which the amino groups of Gly Al and Lys B29 were coupled, using a bifunctional reagent. The groups connecting the two residues were dicarboxylic acids ranging from C2 to C13. All of the derivates had significant in vivo hypoglycemia activity (32 to 100 percent) but markedly reduced in vitro activity on fat cells (2 to 14 percent). Of great in- terest was the observation that NoAl, NeB29 adipoylinsulin could be fully oxidized and then sub- sequently reduced, to give a 75 percent yield of the starting material (5). These data indicate that the 6 carbon bridge is actiong like the connecting peptide of proinsulin to put the chains in the proper position so reduction gives a very high yield of the proper combination of the A and 322 Diabetes Mellitus B chains. Additional research into this kind of molecular modeling may make it possible to com- bine A and B chains with an almost quantitative yield. The development of better methods of chain synthesis and combination would provide for the eventual production of synthetic insulin for treatment of diabetic patients. Eventually, this may become very important as the number of patients with diabetes mellitus is increasing rapidly throughout the world, and we must face the possibility that some day animal sources of insulin may not be sufficient to treat all of the patients who need it. Another benefit of developing better synthetic techniques will be the ability to make large quantities of insulin analogs for potential pharmacologic use, as noted below. B. Long-range Needs: Long range research with insulin analogs has the potential to open new areas of treatment to both diabetic and non-diabetic patients. As noted in the preceeding sections of this monograph, insulin interacts with specific membrane receptors to initiate a wide variety of biological actions. Research to characterize the nature of insulin receptors is essen- tial. Are all insulin receptors the same, or are there different receptors to account for each different type of insulin action? Are the insulin receptors of each insulin-dependent tissue the same, or do they have major differences? Another area which needs extensive investigation is the mechanism(s) by which activation of insulin receptors causes the various biological actions of insulin to occur. Identification of the specific parts of the insulin molecule that interact with insulin receptors needs to be determined. Elucidation of the structure of insulin receptors and the residues of the insulin molecule that interact with them should lead to the synthesis of a variety of simple molecules that could bind to insulin receptors and activate them. This might provide the opportunity for developing parenteral, or oral agents, which have insulin action and could be administered with meals to control the diabetic state. Such a major achievement in therapy would revolutionize the management of patients with diabetes mellitus. As noted in the sections on insulin actions and the relationship of insulin to other anabolic hormones, the insulin molecule has the potential to effect almost every phase of metabo- lism. There are clearly many instances in clinical medicine where a therapeutic agent which had some, but not all of these actions, would be very advantageous. For instance, an agent with the anabolic activities of insulin would be useful in the therapy of growth disturbances, in the repair of injury or in post-operative healing. Likewise, a molecule which had insulin-like action, but did not significantly promote lipogenesis might be extremely useful in the treatment of the obese diabetic patient. The synthesis of insulin analogs and careful evaluation of the spectrum of their biological activities has the potential to find such agents. This depends on whether all of the actions of insulin reside in the same amino acid residues of the molecule. Since the mechanism of insulin action appears to be different for certain activities, it is likely that some insulin analogs will show a spectrum of activity that is different from native insulin. None of the modified or synthetic insulins that have been made in the past have received adequate and careful evaluation of the spectrum of their biological activities. Such studies should be a part of long-range research in this area. Along these same lines of investigation, it would be worthwhile to carefully evaluate the activity of naturally occurring insulins. For example, recent studies have shown that guinea pig insulin has about 10 percent the activity of bovine insulin, when assessed in the mouse convulsion (2.14 1.U.), or rat epidydimal fat, or hemidiaphragm assay, but when assessed by blood glucose Insulin Synthesis and Analogs 323 lowering in the guinea pig, it has 1/3 to 1/2 the activity (70). The relationship of insulin to other anabolic agents, such as somatomedin, nerve growth fac- tor, and epidermal growth factor need to be further clarified. The possibility that insulin analogs might be synthesized, which have increased somatomedin or nerve growth factor activity, should be considered as being highly likely and of great potential therapeutic importance as organ specific growth factors. POTENTIAL VALUE OF THIS RESEARCH The potential values of this research are: (1) the assurance of adequate supplies of insulin for patients with diabetes mellitus; (2) the development of new and better analogs for the treat- ment of diabetes mellitus; (3) the development of new pharmacologic agents for use in diabetes mellitus, and a variety of other diseases. While these goals seem somewhat remote and theoretical, one needs only to look at: the thera- peutic agents that research has made available for use today. We would have been equally skeptical of their development 20 or 30 years ago. REFERENCES 1. Avruch, J, JR Carter, and DB Martin 1972. The effect of insulin on the metabolism of adipose tissue. Handbook of physiology, Section 7, Endocrinology (Vol. 1) Endocrine pancreas. DF Steiner and N Freinkel, eds. Washington, American Physiol Soc pp 545-562. 2. Blundell, TL, GG Dodson, E Dodson, DC Hodgkin, and M Vijayan 1971. X-ray analysis and the structure of insulin. Recent Prog Horm Res 27: 1-40. 3. Blundell, TL, G Dodson, D Hodgkin, and D Mercola 1972. Insulin: The structure in the crystal and its reflection in chemistry and biology. Adv Protein Chem 26:280-402. 4, Brandenburg, D 1972. Preparation of NoAl, NeB29 adipoylinsulin, an intramolecularly cross- linked derivative of beef insulin. Hoppe-Seyler's Z Physiol Chem 353:869-873. 5. Brandenburg, D, and A Wollmer 1973. The effect of a nonpeptide interchain crosslink on the reoxidation of reduced insulin. Hoppe-Seyler's Z Physiol Chem 354:613-627. 6. Brandenburg, D, HG Gattner, M Weinert, L Herbertz, H Zahn, and A Wollmer 1971. Structure function studies with derivatives and analogs of insulin and its chains. Proc. VII Int Diabetes Congress. Excerpta Medica Int Congress, Series 231, pp 363-375. 7. Brandenburg, D, W-D Busse, H-G Gattner, H Zahn, A Wollmer, J Gliemann, and W Puls 1973. Structure-function studies with chemically modified insulins. Peptides. Amsterdam, North- Holland, pp 270-283. 8. Busse, W-D, and H-G Gattner 1973. Selective cleavage of one disulfide bond in insulin: Preparation and properties of insulin A7-B7-di-S-sulfonate. Hoppe Seyler's Z Physiol Chem 354:147-155. v 9. Butcher, RW, GA Robison, and EW Sutherland 1972. Biochemical actions of hormones (Vol. II). New York, Academic Press. pp 21-54. 10. Carpenter, FH 1966. Relationship of structure to biological activity of insulin as revealed by degradative studies. Amer J Med 40:750-758. 11. Cautrecasas, P 1972. Insulin actions. New York, Academic Press. pp 137-163. 12. Daughaday, WH, and JT Garland 1972. Growth and growth hormone. Amsterdam, Excerpta Medica Foundation, pp 168-179. 324 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. zl. Diabetes Mellitus De La Haba, G, GW Cooper, and V Elting 1966. Hormonal requirements for myogenesis of striated muscle in vitro: Insulin and somatotropin. Proc Nat Acad Sci (USA) 56:1719-1723. Du, Y-C, R-Q Jiang, and C-L Tsou 1965. Conditions for successful resynthesis of insulin from its glycyl and phenylalanyl chains. Sci Sinica, Peking 14:229-236. Fain, JN, and L Rosenberg 1972. Antilipolytic action of insulin on fat cells. Diabetes 21 (suppl. 2):414-425, Frazier, WA, RH Angeletti, and RA Bradshaw 1972. Nerve growth factor and insulin. Science 176:482-488. Fritz, IB 1972. Insulin action. New York, Academic Press, pp 571-602. Hackel, DB, E Mikat, HE Lebovitz, and K Schmidt-Nielsen 1967. Diabetes mellitus-like disease in sand rats (Psammomys Obesus). Proc 6th Int Cong Diabetes. Excerpta Medica Series 172: 800-805. Hall, K, and K Uthne 1971. Some biological properties of purified sulfation factor from human plasma. Acta Med Scand 190:137-143. Hintz, RL, DR Clemmons, L Underwood, and JJ Van Wyk 1972. Competitive binding of somatomedin to the insulin receptors of adipocytes, chondrocytes, and liver membranes. Proc Nat Acad Sci (USA) 69:2351-2353. Hodgkin, DC, and D Mercola 1972. The secondary and tertiary structure of insulin. Handbook of physiology, Section 7, Endocrinology (Vol. 1) Endocrine pancreas. Washington, American Physiol Soc pp 139-157. Hornle, S, U Weber, and G Weitzel 1968. Struktur und wirkung von insulin IV synthetische A-ketten mit variierten sequenz. Hoppe Seylers Z Physiol Chem 349:1428-1430. Humbel, RE, HR Bosshard, and H Zahn 1972. Chemistry of insulin. Handbook of physiology, Section 7, Endocrinology (Vol. 1) Endocrine pancreas. Washington, American Physiol Soc pp 111-132, Jost, K, and J Rudinger 1968. Synthesis of peptides corresponding to sequences from chain A of sheep insulin, with cystine replaced by cystathione. Collection Czech Chem Commun 33:109-118. Jost, K, J Rudinger, H Klostermeyer, and H Zahn 1968. Synthese und hypoglycamische wirkung eines insulinanalogen cystathionin - peptides: ein argument gegen die beteiligung der ‘intra- chenaren disulfid-gruppe bei der insulinwirkung. Z Naturforsch 236:1059-1061. Katsoyannis, PG 1969. Synthetic insulins. Diabetes: Proc. VI Congress IDF Excerpta Medica Foundation International Congress Series 172. pp 379-394. Katsoyannis, PG, and A Tometsko 1966. Insulin synthesis by recombination of A and B chains: A highly efficient method. Proc Nat Acad Sci (USA) 55:1554-1561. Kipnis, DM 1970. Insulin secretion in normal and diabetic individuals. Advances in internal medicine (Vol. 16) Year Book Publishers. pp 103-134, Krahl, MD 1972. Effects of insulin on synthesis of specific enzymes in various tissues. Insulin action, New York, Academic Press. pp 461-486. Larner, J, C Villar-Pilasi, NO Goldberg, JS Bishop, F Huiijing, JI Wenger, H Sasko, and NB Brown 1968. Hormonal and nonhormonal control of glycogen synthesis--Control of transferase phosphatase and transferase I kinase. Advan Enzyme Regul 6:409-423, Lebovitz, HE, and JM Feldman 1973. Pancreatic biogenic amines and insulin secretion in health and disease. Fed Proc 32:1797-1802. 32. 33, 34. 35. 36. 37. 38. 29. 40. 41. 42, 43, 44, 45, 46. 47. 48. 49. 50. Insulin Synthesis and Analogs 325 Levy, D, and FH Carpenter 1970. Insulin methyl ester. Specific cleavage of a peptide chain resulting from a nitrogen to oxygen acyl shift at a threonine residue. Biochemistry 9:3215-3222, Lowenstein, JM 1972, Is insulin involved in regulating the rate of fatty acid synthesis. Handbook of physiology, Section 7, endrocrinology (Vol. 1) Endocrine pancreas. Washington, Amer Physiol Soc, pp 415-424, Lubke, K, and H Klostermeyer 1970. Synthese des insulins, anfinge und fortschritte. Adv Enzymol 33:445-525. Madison, LL 1969. Role of insulin in the hepatic handling of glucose. Arch Intern Med 123:284-292, Manchester, KL 1970. Insulin and protein synthesis. Biochemical action of hormones (Vol 1) New York, Academic Press pp 267-320. Manchester, KL 1972. Effect of insulin on protein synthesis. Diabetes 21: (suppl. 2) 447-452, Morgan, HE, DE Rannels, EB Wolpert, KE Giger, JW Robertson, and LS Jefferson 1972, Effect of insulin on protein turnover in heart and skeletal muscle. Insulin action. New York, Academic Press, pp 437-460. Morris, JWS, DA Mercola, and EF Arquilla 1970. Preparation and properties of 3-nitrotyrosine insulins. Biochemistry 9:3930-3937. ) Park, CR, JGT Sneyd, JD Corbin, LS Jefferson, and JH Exton 1969. Role of cyclic adenylate in the actions of insulin. Diabetes: Proc VI Cong IDF. Excerpta Medica Foundation Int Cong Series 172, pp 5-15. Park, CR, SB Lewis, and JH Exton 1972. Relationship of some hepatic actions of insulin to the intracellular level of cyclic adenylate. Insulin action. New York, Academic Press, pp 509-528. Peck, WV, and K Messinger 1970. Nucleoside and ribonucleic acid metabolism in isolated bone cells. Effects of insulin and cortisol in vitro. J Biol Chem 245:2722-2729, Podolsky, S 1971. Lipoatrophic diabetes and miscellaneous conditions related to diabetes mellitus. Joslin's Diabetes Mellitus 11th Edition, Philadelphia, Lea and Febiger, pp 722-733. Rager, K, W Kemmler, and P Schauder 1969. Biologische und immunologische aktivitat chemisch abgewandelter insuline. Hoppe Seyler's Z Physiol Chem 350:717-720. Riggs, T 1970. Hormones and transport across cell membranes. Biochemical actions of hormones. New York, Academic Press, pp 157-208. Ryle, AP, F Sanger, LF Smith, and R Kitai 1955. The disulphide bonds of insulin. Biochem J 60:541-556. Salmon, WD, Jr, MR Bavall,, gnd EY Thompson 1968. Stimulation by insulin in vitro of incorpora- tion of (35s sulfate and (1%C) leucine into protein polysaccharide complexes, (3H)uridine into RNA and (3H)thymidine into DNA of costal cartilage from hypophysectomized rats. Endocrinology 82:493-499. Sanger, F 1945. The free amino groups of insulin. Biochem J 39:507-515. Stockdale, PE, and YJ Topper 1966. Role of DNA synthesis and mitosis in hormone dependent differentiations. Proc Nat Acad Sci (USA) 56:1283-1289. Taylor, JM, S Cohen, and WM Mitchell 1970. Epidermal growth factor: High and low molecular weight forms. Proc Nat Acad Sci (USA) 67:164-171. 326 51. 52. 53. 54. 55. 56. 57. 58. BY. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. Diabetes Mellitus Underwood, LE, RL Hintz, SJ Voina, and JJ VanWyk 1972. Human somatomedin, the growth hor- mone dependent sulfation factor, is anti-lipolytic. J Clin Endocrinol Metab 35:194-198, Weber, U, and G Weitzel 1968. Struktur und wirkung von insulin V synthetische B-ketten mit variierter sequenz. Hoppe Seyler's Z Physiol Chem 349:1431-1433. Weber, U, F Schneider, P Kohler, and G Weitzel 1967a. Struktur und wirkung von insulin I synthetische A-Ketten mit variierter sequenz. Hoppe Seyler's Z Physiol Chem 348:947-949. Weber, U, G Hoérnle, G Grieser, K-H Herzce, and G Weitzel 1967b. Struktur und wirkung von insulin II synthetische A-ketten mit variierter sequenz. Hoppe Seyler's Z Physiol Chem 348:1715-1717. Weber, U, S Hornle, P Kohler, G Nagelschneider, K Eisele, and G Weitzel 1968. Struktur und wirkung von insulin III synthetische A-ketten mit variieter sequenz. Hoppe Seyler's Z Physiol Chem 349:512-514. Weitzel, G, U Weber, S Hornle, and F Schneider 1968. Struktur und wirkung von insulin: Synthetische A-Ketten mit variierter sequenz. Peptides. Proc IX European Peptide Symposium. Amsterdam, North Holland pp 222-227. Weitzel, G, K Eisele, H Zollner, and U Weber 1969. Struktur und wirkung von insulin VII verkirzte synthetische B-ketten. Hoppe Seyler's Z Physiol Chem 350:1480-1483. Weitzel, G, U Weber, K Eisele, H Zollner, and J Martin 1970. Struktur und wirkung von insulin VII austausch von histidin gegen alanin in synthetischen B-ketten. Hoppe Seyler's Z Physiol Chem 351:263-267. Weitzel, G, U Weber, J Martin, and K Eisele 1971. Struktur und wirkung von insulin X beteiligung von arginin B22 an der insulinwirkung. Hoppe Seyler's Z Physiol Chem 352:1005- 1013. Wool, IG 1969. Action of insulin on protein synthesis. Diabetes: Proc VI Cong IDF. Excerpta Medica Foundation Int Cong Series 172, pp 16-24. Wool, IG, JJ Castles, DP Leader, and A Fox 1972a. Insulin in the function of muscle ribosomes. Handbook of physiology Section 7 Endocrinology (Vol. 1) Endocrine pancreas. Washington, Amer Physiol Soc, pp 385-394. Wool, IG, REH Wettenhall, H Klein-Bremhaar, and N Abayang 1972b. Insulin and the control of protein synthesis in muscle. Insulin action. New York, Academic Press, pp 415-424. Wrenshall, GA, B Bogoch, and RC Ritchie 1952. Extractable insulin of pancreas. Diabetes 1:87-107. Yalow, RS, and SA Berson 1960. Immunoassay of endogenous plasma insulin in man. J Clin Invest 39:1157-1175. Zahn, H, B Gutte, EP Pfeiffer, and J Ammon 1966. Resynthese von insulin aus praoxydierter A-kette und reduzierter B-kette. Ann Chem 691:225-231, Zahn, H, and E Drechsel 1968. Resynthese von insulin aus A-ketten monodisulfid-bis-sulfonas. Hoppe Seyler's Z Physiol Chem 349:385-389. Zahn, H, W Danho, G Schmidt, K Jost, H Klostermeyer, E Drechsel, and HG Gattner 1969. Pro- gress in the synthesis of insulin and new A- and B-chain compounds. Diabetes: Proc VI Cong IDF. Excerpta Medica Foundation Int Cong Series 172, pp 367-378. Zahn, H, D Brandenburg, and H-G Gattner 1972. Molecular basis of insulin action: Contribu- tions of chemical modifications and synthetic approaches. Diabetes 21(suppl 2): 468-475, Zierler, KL 1972. Insulin, ions, and membrane potentials. Handbook of physiology (Vol. 1) Endocrine pancreas. Washington, Amer Physiol Soc, pp 347-368. Zimmerman, AE, DIC Kellis, and CC Yip 1972. Physical and biological properties of guinea pig insulin. Biochem Biophys“Res Comm 46:2127-2133. DRUGS ENHANCING INSULIN SECRETION Harold E. Lebovitz INTRODUCTION [The main objective in the treatment of patients with diabetes mellitus is to restore their metabolic balance to normal | We assume that goal will be achieved if the diabetic patient's blood glucose is maintained at the same fasting and postprandial levels as that of normal individuals. This will be a valid assumption if the agents used for treatment have the same spectrum of bio- logical activities as insulin, and the defects being treated can be attributed to insulin in- sufficiency. (at diabetic patients have a total lack of pancreatic insulin (destruction of islet cells or disturbance of insulin synthesis) therapy must consist of the administration of exogenous insulin or insulin-related peptides. If diabetic patients have adequate stores of pancreatic insulin but are unable to secrete it appropriately because of a disturbance in the secretory process, it should be possible to treat those individuals with agents that stimulate the release of insulin. If diabetic patients have adequate or excessive quantities of pancreatic or circulating insulin but its actions are blocked at the cellular level, then therapy must be directed toward administering massive doses of exogenous insulin or modifying the cellular response to the endogenous insulin. | The present chapter explores the development and potential use of agents that increase insulin secretion. We shall first try to define the patients who might benefit from such a therapeutic ap- proach. Next we will review the current state of knowledge about the mechanisms of insulin secre- tion. We will then examine the possibility that some forms of diabetes mellitus may be due to specific defects in the insulin secretory process. And, finally, we will define the potential approaches that might be taken to develop agents that will stimulate insulin secretion and may therefore be useful in the treatment of some patients with diabetes mellitus. EVIDENCE THAT ABNORMALITIES OF INSULIN SECRETION ARE IMPORTANT IN THE PATHOGENESIS OF DIABETES MELLITUS Diabetes mellitus is not a single disease, but rather a group of diseases that are similar in that the resultant metabolic defect is due to insufficient insulin action. Diabetes mellitus in man can be subdivided as shown in Table 1. Spontaneous genetic diabetes mellitus is itself a com- plex disease which exists in two major and possibly many minor variants. The major variants, ketotic insulin @ependent diabetes mellitus and nonketotic maturity-onset diabetes mellitus ex- hibit many striking differences (64). The ketotic insulin dependent form ordinarily manifests itself prior to 25 years of age. The pancreas contains little or no insulin. (97). Plasma insulin levels are low and are unchanged by the administration of glucose or other known stimulators of insulin secretion (74). These patients represent approximately 5 percent of the total diabetic population (63,65). They require exogenous insulin administration for their treatment. Non- ketotic maturity-onset diabetes mellitus ordinarily presents after the age of 25 years and is usually associated with obesity, multiparity or some other predisposing environmental factor (63, 65). The insulin content of the pancreas is reduced (mean 50 percent with a range of 20 to 100 percent) compared to matched controls but is still about 1 U/g (mean) (97). Fasting plasma { . 227 328 Diabetes Mellitus TABLE 1. Types of Diabetes Mellitus TABLE 2. Physiologic Factors that Stimulate ie . In Vitro Insulin Secretion 1. Spontaneous Genetic Diabetes Mellitus 1. Carbohydrates a. Ketotic Insulin Dependent b. Maturity Onset (Nonketotic) a. Glucose b. Mannose and other Metabolizable 2. Endocrine Induced Monoszecharides a. Acromegaly b. Cushing's Syndrome 2. Hormones c. Hyperthyroidism a. B Adrenergic Receptor Stimulators d. Pheochromocytoma b. Glucagon e. Carcinoid Syndrome c. ACTH : 3 d. TSH 3. Destruction of Pancreatic Islet Cells ec. Castro-intestinal Hormones 2 I 1. Secretin ; . in-Chol tokinin (? c. Inflammation (Pancreatitis) : a Cholecys *) <, jamr Infiton tosi 4. Enteroglucagons e. Metabolic (Hemochromatosis) 5. VIP (Vasoactive Intestinal Peptide) 4. Resistance to Insulin Action 6. GIP (Gastric Inhibiting Peptide) a. Lipoatrophic Diabetes Mellitus f. Cyclic AMP b. Werner's Syndrome Aino: Acide 4. Cholinergic Agents 5. Cations a. Increase Flux of Sodium Out of the Cell b. Increased Extracellular Potassium c. Increased Intracellular Calcium insulin levels are moderately elevated and the plasma insulin response to oral glucose or meal ingestion is either normal or exaggerated (98). Initially it was thought that this indicated that the primary defect in maturity-onset diabetes mellitus was peripheral resistance to insulin action. More extensive studies, however, indicated that this is not so, for if the blood glucose levels are taken into account and appropriate studies are done to match blood glucose curves in maturity-onset diabetic patients and normal individuals, it is clear that for a given blood glu- cose curve, maturity-onset diabetic patients secrete less insulin than do normal individuals (43,75,76) . In addition, a very striking characteristic of glucose-mediated insulin secretion in maturity-onset diabetic patients is a delay in both the onset of secretion and the time at which peak secretion occurs relative to the time at which the plasma glucose peaks (98). Studies mea- suring the plasma insulin response to intravenously administered glucose in maturity-onset diabetic patients as compared to age, weight, and sex matched controls confirm the delayed and impaired insulin secretion (88). Maturity-onset diabetic patients comprise about 90 to 95 percent of the diabetic population. Additional support for the concept that impaired insulin secretion is characteristic of maturity-onset diabetes mellitus has come from several studies in which glucose= mediated insulin secretion was shown to be impaired in pre-diabetic patients with normal glucose tolerance (3,10,92). It seems reasonable to conclude that a major defect in maturity-onset diabetes mellitus is one or more abnormalities in insulin secretion. Thus agents which stimulate the secretion of insulin might be expected to be useful in the treatment of this type of diabetes and an understand- ing of the exact nature of the defect might lead to development of drugs specific for the particu- lar defect. Drugs Enhancing Insulin Secretion 329 The other forms of diabetes mellitus listed in Table 1 are clearly due to lack of pancreatic insulin oi" peripheral resistance to insulin action. Endocrine induced diabetes mellitus probably representess a combination of impaired insulin release and peripheral antagonism of insulin action. Drugs stimu lating insulin secretion might be expected to be useful in these disorders, particularly if, as has been suggested by several investigators, diabetes is induced only in that 15 or 20 percent of tlie population that may carry some of the genes for spontaneous diabetes mellitus (9, 12,13). MODEL FOR INSUL.IN SECRETION Insulin secretion is a complex process. It has been discussed in some detail in this mono- graph. In order to understand the development of drugs to enhance insulin secretion it is, however, necessary to review some of the more important aspects of current models of insulin secretion. [Insulin secre tion may be thought of as a three-step process which involves (1) recognition of the insulinogeni ¢ stimulus .by the beta cell, (2) generation of some appropriate intracellular message, and (3) act'ivation of a granule releasing system. Disturbances of insulin secretion could occur through abnormalities in any of these components and drugs that stimulate insulin secretion might do so at any of these sites. (1) Recognition of the insulinogenic stimulus by the beta cell Insulin secretion by the beta cell is stimulated by many agents. Table 2 lists some of the more important physiolog.ic factors that directly stimulate insulin secretion in vitro. The most important physiologic age'nt which stimulates insulin secretion is glucose. Some evidence suggests that the signal for gluco.se-mediated insulin release is an as yet unknown phosphorylated metabolite of glucose. Sugars that a re readily metabolized (i.e., glucose, mannose, and to a lesser extent, fructose) stimulate insuliin release(16,17,38). Sugars that are not metabolized (i.e., galactose, 3-0-methyl glucose) do not stimulate insulin release (17,38). D-mannoheptulose, which is an inhibitor of glucose and mar nose uptake) inhibits insulin secretion stimulated by carbohydrates, but not by other agents (16, 17,18). Matschinsky et al. (66,67) have been unable to show any alterations in beta cell gluc ‘ose metabolites that could account for glucose-mediated insulin secretion, and they suggest t hat glucose stimulates insulin release by directly interacting with a plasma membrane glucose receptor. All hormones which stimulate in vitro insulin secretion probably do so by interacting vith a specific plasma membrane receptor, activating beta cell adenylate cyclase and generatin g increased intracellular cyclic AMP (55,84). It is unlikely that ACTH and TSH are significant fac 'tors in controlling insulin secretion. Whether alpha cell pan- creatic glucagon influences beta cell insulin secretion through intercellular bridges in the islets is unknown. Recent information : suggests that the gastrointestinal hormones may be quite important in the physiologic regulation of insulin secretion in response to oral nutrients (2,6,26). The physiologic role of the beta receptor is unclear (53). Butyrylated derivatives of cyclic AMP and drugs that inhibit adenosine 3 “5”-monophosphate diesterase (such as theophylline and caffeine) are potent stimulators of insulin : secretion (30,55). It is of particular importance, however, to note that cyclic AMP and theophy lline can only potentiate glucose-mediated insulin secretion. In the total absence of medium and i intracellular glucose, they are unable to stimulate insulin secretion(5,31). Therefore, cyclic AMP and hormones which act through it must be viewed as agents that potentiate glucose-mediated inst .ilin secretion and not as primary stimulators of secretion. Amino acids may stimulate imsulin sec retion through several different mechanisms (26): (1) direct 330 Diabetes Mellitus interaction with a plasma membrane receptor; (2) transport across the cell membrane; (3) indirectly by releasing adjacent alpha cell glucagon which stimulates beta cell insulin secretion; (47) through metabolites formed intracellularly. Leucine stimulates insulin secretion in the absence Of glucose (69) and an unmetabolizable analogue of leucine (BCH), also stimulates insulin secretion (28). These data suggest that leucine itself stimulates insulin secretion. Arginine stiimulation of insulin secretion is mediated differently from leucine [diazoxide inhibits leucine-ir.duced, but not arginine-induced insulin secretion; mannoheptulose suppresses arginine but not leuc.ine-induced insulin secretion (26)]. Cholinergic agents potentiate glucose mediated insulin release through a muscarinic action that is blocked by atropine (52,56,89), Cations are not specifical’ly recognized by the beta cell but are involved in either the recognition of insulinogenic stimuli or generation of an intracellular messenger (53,84). The beta cell therefore recognizes physiologic stimulators of insulin secreticn by a variety of mechanisms which include cell membrane receptors, transport across the cell mer brane, or the production of a metabolite. (2) Generation of an appropriate intracellar message Following the recognition of an insulinogenic stimulus, the beta cell must have one or more mechanisms to transfer this message to the granule releasing system. Evidence for two purportedly interrelated systems exist. The first is an ionic shift in sodium and calciurp; the second is the beta cell cyclic AMP system. Malaisse (54) has characterized agents that stimulate in vitro insulin secretion in relation- ship to their effects as the concentration of glucose in the medium is incr eased from none to 750 mg/dl (Fig. 1). Agents which lower the "Km" without affecting the "Vma.x" of the secretory process are classified as glucose-stimulating agents. Agents which do not: affect the "Km" but increase the "Vmax" of the secretory process are referred to as glucose-pc tentiating agents. Glucose simulating agents are glucose, other sugars, certain amino acids, such as leucine and different lipid metabolites. Glucose-potentiating agents are adrenergic and polypeptide hormones that act through the adenylate cyclase system. Dean and Matthews (19,20,21) showed that stimulation of insulin re:.’lease in the beta cell is accompanied by depolarization of the beta cell membrane. Glucose stimu lation of impaled mouse beta cells decreased the membrane potential from -33mV to -16mV. Associated with increasing glucose concentrations were action potentials 1-4mV in amplitude and ciccurring in bursts with an interval of three to four seconds. Other insulinogenic stimuli prodiux:ed similar electrical activity. These electrical changes seem to be caused by calcium entering into t.he beta cell and exchanging with intracellular sodium. Malaisse in a separate series of papers (54,57,58) has prese:nted evidence that glucose- stimulating agents increase the net uptake of calcium by the beta « ell. This appears to be due to a reduction in efflux of calcium. Glucose-potentiating agents (5, £,4) were shown to increase the efflux of Ca“® from the cell without altering the uptake of calciurm. He has hypothesized that insulin secretion is stimulated by increases in the cytosol calcium concentration. Calcium is thought to be in two intracellular pools: the cytosol and within (cell organelles. Glucose- potentiating agents are purported to translocate large quantities; of calcium from the organelles to the cytosol, thereby increasing cytosol levels. The increased CaS efflux is postulated to re- flect the movement of some of this increased calcium to the extr acellular space. Glucose-stimulating TOTAL INSULIN OUTPUT +100 INCREMENT IN OUTPUT *—=+GLUCOSE (200mg/ml) ~—+GLUCOSE (100 mg/ml) GLUCOSE SIMULATION ~~ GLUCOSE EFFECT x 20 ~~ GLUCOSE EFFECT x1.5 GLUCOSE POTENTIATION E E + 50 & XA a ol ol a @ 0.2 500 750 0 200 500 E 100 300 100 2 250; - = a 200f = o z 150 | fe +150 — 3 Z 100t +100 rr 50+ + 50 a OL 1 1 1 1 J OL 1 L J 0 200 500 750 0 200 500 100° 300 100° 300 GLUCOSE CONCENTRATION (mg/100ml) agents are thought to increase cytosol calcium concentrations by decreasing calcium efflux. Drugs Enhancing Insulin Secretion 331 FIGURE 1. Definition of glucose-stimu- lating (upper panel) and glucose-poten- tiating (lower panel) insulinotropic agents. The total insulin output (left) and the increment in secretion rate (right) evoked by these agents in pieces of rat pancreatic tissue are shown as a function of the glucose concentration of the incubation medium (indicated along the abscissa). On the left, the rate of insulin output induced by glucose alone is shown by the heavy line. The other curves were derived from the experi- mental data, as shown in the rectangles, by assuming either that a fixed amount of glucose (1.0 or 2.0 mg/ml) was added to the glucose already present in the incubation medium (glucose stimula- tion), or that the insulinotropic cation of glucose was multiplied by a constant factor (x1.5 or 2.0; glucose potentiation). From Malaisse, W. J. 1973. Diabetologia 9:167-173. Reprinted with permission from Springer-Verlag. The control of insulin secretion by increasing cytosol calcium concentration is a provoking theory for which there is some circumstantial evidence. One must, however, note that such a theory in which a sequestered intracellular calcium pool is hypothesized to interact with an active cytosol pool that is also interacting with the external environment is not presently amenable to testing. Direct measurements of the cytosol calcium concentrations under various states of insulin secre- tion will be necessary to prove this theory. The recognition of many agents which increase insulin secretion is accompanied by an increase in intracellular cyclic AMP. Glucagon, theophylline, and glucose have all been shown to increase islet cell cyclic AMP content (15,44,93). Cyclic AMP and theophylline were shown to increase the efflux of CaS from prelabeled isolated islets and, as noted above, this has been interpreted to mean that intracellular cyclic AMP increases cytosol calcium by translocating calcium from cellular organelles to cytosol. It is not presently known whether the increase in intracellular cyclic AMP mediates other intracellular effects such as phosphorylating proteins or activating gene action. It is also uncertain whether there are additional intracellular messengers, generated by one or more of the various agents, that stimulate insulin secretion. 332 Diabetes Mellitus 3. Activation of granule releasing system Anatomical studies coupled with insulin secretory studies have suggested the presence of an intracellular transport system for beta granules (45,47). Insulin is synthesized as proinsulin within the endoplasmic reticulum of the beta cell. An energy dependent process transfers it to the Golgi apparatus where distinct beta cell granules are formed and released into the cytoplasm. The granule is surrounded by a smooth membranous sac. Proinsulin is cleaved into insulin and con- necting peptide as the granule is formed. Zinc is incorporated into the insulin molecule in the mature beta granule. The beta granules in the cytoplasm attach to microtubules which have a diameter of 200A and a wall comprised of 12 to 14 subunits. The microtubules run perpendicular to the plasma membrane. The beta granules appear to migrate down the microtubules to the plasma membrane. The exact mechanism of movement is unknown. On the inner side of the plasma membrane is a layer of short interconnected fibers which are compact and exclude most of the other com- ponents of the cell. This region, called the cell web, is composed of microfilaments (approxi- mately 40 to 70 A°) and appears to be attached to the inner surface of the plasma membrane (59, 73). It creates a zone of exclusion through which the beta granules have to pass on their way to the plasma membrane. When the beta granule reaches the plasma membrane, its membrane fuses with the plasma membrane and both dissolve at the point of contact and discharge the granule's content outside of the cell with formation of microvillous projections at the discharge site. This process is called emiocytosis (47). Recently, the outside of the beta cell plasma membrane has been shown to be coated with a carbohydrate layer composed of glycoproteins and mucopolysaccharides (73). The function of this outer layer is unknown. Agents which stimulate insulin release appear to do so by causing movement of granules down the microtubular system to the cell surface where emiocytosis takes place(47,48). The evi- dence that the microtubular system is involved in insulin secretion is: (1) drugs that destroy microtubules such as colchicine, vincristine, and vinblastine inhibit insulin secretion stimulated by all agents (glucose, leucine, sulfonylurea drugs, etc.); (2) chemicals that stabilize micro- tubules such a deuterium oxide, hexylene glycol (0.15 to 1.0 percent) or ethanol (1.0 percent) cause reversible inhibition of insulin secretion stimulated by glucose, leucine, or sulfonylurea drugs; (3) these antimicrotubular agents inhibit both the initial and late phase of glucose- stimulated insulin release from perifused isolated islets. The role of the microfilamentous cell web is not as clearly defined. The effect of cytochalasin B on the ultrastructure of the beta cell web and insulin secretion by isolated islets has been studied by Orci et al. (72) and Malaisse et al. (59). Cytochalasin B markedly disrupted the cell web of isolated islets and though it had no effect on basal insulin release, it markedly enhanced glucose-stimulated insulin secretion. This enhancement was reversible and could be inhibited by deuterium oxide. In perifused systems, cytochalasin B enhanced both the early and late phase of glucose-mediated insulin release (49). These studies suggest that the microfilamentous web may serve as a means of deterring insulin secretion. Recent studies with cytochalasin B, however, indicate that this drug inhibits the uptake of glucose, glucosamine and 2-deoxy-D-glucose in HeLa cells and pancreatic islets and also inhibits glycoprotein and mucopolysaccharide synthesis in embryonic cells (68,87). Therefore it is possible that the effects of cytochalasin B are not on the microfilamentous systems but on the cell plasma membrane, or some other component of the cell. Drugs Enhancing Insulin Secretion 333 Evidence for emiocytosis as the end process of secretion has been amply documented by both scanning election microscopy and freeze-fracture electron microscopy (46,73). The granule releasing system is activated by intracellular messages generated by the insulin releasing stimuli. Studies by Grodsky and Bennett (37) and Milner and Hales (70) showed that in- sulin secretion in vitro could not occur in a calcium free medium regardless of the stimulating agent. Their data and the data discussed above, on Ca*® uptake and efflux, have led many investi- gators to hypothesize that changes in calcium ion trigger the beta cell microtubular system to transport and release beta granules. The mechanism by which calcium ions trigger the microtubular- microfilamentous system is not clear. The mechanism of intracellular cyclic AMP in activating the microtubular-microfilamentous system is also unclear. As noted above, Malaisse has hypothesized that cyclic AMP increases cytosol calcium ion and that its effects are therefore calcium mediated. Other models of cyclic AMP action have supported this concept. Another possibility, however, is that cyclic AMP may activate a protein kinase that phosphorylates some protein which effects granule transport. In addition to physiologic agents that stimulate insulin secretion, there are several physio- logic factors which are known to inhibit insulin secretion. These are listed in Table 3. TABLE 3. Physiologic Factors that Inhibit Insulin Secretion 1. ao Adrenergic Receptor Stimulators 2. Dopamine 3. Serotonin 4. Amonium Ion Alpha adrenergic receptor stimulating agents inhibit both in vivo and in vitro insulin secretion (80). They do so through interference with the discharge of beta granules. Turtle and Kipnis (93) indi- cated that they do so by inhibiting beta cell adenylate cyclase and depressing intracellular cyclic AMP levels. Feldman and Lebovitz (30) have shown that epinephrine and other alpha adrenergic stimulating agents (32) block the in vitro insulin secretory response to dibutyryl cyclic AMP and suggested that epinephrine interferes directly with the granule release process. Brisson and Malaisse (4) have tried to explain epinephrine inhibition on insulin secretion through a decrease in cytosol calcium. Their studies show very small and transient effects of epinephrine in in- creasing Ca"® efflux from the beta cell and they hypothesize that epinephrine provokes a trans- location of calcium from the cytosol to some organelles. The data are not very convincing. Dopamine and serotonin are reported to occur in the cytosol of beta cells of some species (8). Studies reported have indicated that both of these monoamines interfere with the final stages of granule secretion. Lebovitz and Feldman (51) have proposed that serotonin and/or dopamine occur in or near the beta granule and tonically inhibit migration down the microtubule. Secretion is thought to be the net balance between stimulatory influences and the tonic inhibition by these monoamines. Serotonin antagonists markedly potentiate glucose and tolbutamide stimulated insulin secretion (33). It is not known whether these monoamines inhibit insulin secretion by altering local calcium concentrations or interfering with microtubular function. The ammonium ion interferes with insulin secretion through altering either the recognition of glucose dependent stimulators or the generation of their intracellular message. It does not interfere with the granule secretory system since it does not block sulfonylurea drug-mediated insulin release (31). 334 Diabetes Mellitus NATURE OF THE INSULIN SECRETORY DEFECT IN MATURITY-ONSET DIABETES MELLITUS Many studies have been done to try to characterize the nature of the defect in insulin secre- tion in patients with diabetes mellitus. Difficulties have been encountered in these studies and much of the data are controversial. The problems encountered are related to several factors: (1) age, sex, diet, and degree of obesity influence the insulin secretory response; (2) severity of diabetes and the quantity of insulin remaining in the pancreatic beta cell will affect the magni- tude of the insulin secretory response; (3) most clinical studies utilize plasma insulin changes as the index of secretory response; (4) all clinical studies measure the effects of the stimu- lating agents in the presence of normal or elevated extracellular glucose. In spite of the above difficulties, relevant information is available on the effect of a num- ber of agents on insulin secretion in maturity-onset diabetic patients and prediabetic individuals. 1l. Glucose Many different techniques have been used to study the insulin response to glucose. Oral administration of glucose to patients with mild or moderate maturity-onset diabetes is associated with the following plasma insulin changes: (1) the onset of the rise in plasma insulin is delayed as compared to that occurring in normal individuals, (2) the peak plasma insulin is usually higher than in normals and occurs after the peak plasma glucose is reached, (3) total insulin secretion is frequently greater than that seen in normals (98). Because diabetic patients extract less oral glucose in the liver than normal individuals, their plasma glucoses are much higher than normals. Thus, it is difficult to compare insulin secretion following oral glucose in diabetics to normals since the glycemic stimulus is greater. Perley and Kipnis (75,76) attempted to solve this dilemma by simulating the oral plasma glucose curves with computer programmed glucose infusions. Maturity- onset diabetics had lower plasma insulins than normals with both normal and diabetic simulated glucose tolerance curves. Seltzer et al. (88) attempted to solve this problem by calculating plasma insulin changes relative to, plasma glucose changes and expressing it as an index. This calcula- tion also indicated impaired insulin secretion following oral glucose in diabetic patients. The plasma insulin response to oral glucose has been studied in prediabetic patients (in which, of course, glucose tolerance is normal) and are reported to be normal or decreased (27,339,382). Studies of the plasma insulin response to intravenous glucose have uniformly demonstrated that maturity-onset diabetic patients and prediabetic individuals have impaired insulin secretion in response to glucose (3,88). Most of these studies have used a 25 g glucose bolus as the stimu- lant. Cerasi and Luft (9,10) have developed an intravenous glucose procedure to measure insulin secretion. They give a priming intravenous injection of glucose followed by a constant infusion for 60 minutes. Plasma glucose and insulin are determined frequently and the insulin response analyzed in relation to the glucose stimulation by an analogue computer. Using this technique they have defined a parameter known as the initial insulin response. They found that this initial response is markedly decreased in patients with maturity-onset diabetes, prediabetic patients, and 15 to 20 percent of normal controls. They suggest that this impaired initial secretory response is a marker of the genetic abnormality of diabetes mellitus. Defective insulin secretion in response to glucose stimulation appears to be one of the key defects in maturity-onset diabetes mellitus. Drugs Enhancing Insulin Secretion 335 2. Amino Acids Amino acids alone or in mixtures stimulate insulin secretion in man when they are given orally or intravenously. Patients with maturity-onset diabetes mellitus and subclinical diabetes show lower than normal rises in plasma insulin levels during intravenous infusions of amino acids (26,28,35). The plasma insulin response to intravenous amino acids in prediabetic subjects is normal. 3. Beta Receptor Stimulators Intravenous administration of beta adrenergic agents such as isoproterenol increase the plasma insulin of normal humans (79). Several studies have attempted to implicate a defective pancreatic beta cell beta receptor as the cause of the impaired insulin secretion in diabetes mellitus (11,14), Specifically, infusions of propranolol are reported to significantly inhibit glucose-mediated insulin release. Deckert and colleagues (24), however, showed that isoprenaline infusion stimulated insulin secretion in maturity-onset diabetics just as well as in normals. Robertson and Porte (85) showed that maturity-onset diabetic patients showed the same magnitude of insulin secretion in response to isoproterenol as normals, even though they had markedly im- paired insulin secretion in response to intravenous glucose. Isoproterenol, but not glucose- stimulated insulin secretion, could be blocked by propranolol. They concluded that the glucose receptor is distinct from the beta receptor and that the latter is not involved in the insulin secretory defect in diabetes mellitus. 4. Alpha Receptor Stimulators Intravenous administration of alpha adrenergic agonists in man inhibit insulin secretion (80). Agents that block the alpha adrenergic receptor stimulate insulin secretion in normals and maturity-onset diabetics (77.80). The administration of alpha adrenergic receptor antagonists (phentolamine) increase the insulin secretory response to intravenous glucose in normal subjects (7). Of considerable interest is the observation that even though insulin secretion was greater, the glucose disposal constant was unchanged. Similarly, administration of phentolamine to healthy fasting volunteers caused hyperinsulinemia in response to an intravenous glucose load without amelioration of the glucose intolerance (71). Efendic, Cerasi, and Luft (25) reported that blockage of alpha adrenergic receptors partially restores the initial insulin response (to glucose) in their prediabetic subjects toward normal. They claim that a similar treatment had no effect on early secretion in normal subjects. 5. Secretin and Other GI Hormones Deckert (22) showed that 75 units of secretin intravenously elicited the same striking rise in plasma insulin in eight maturity-onset diabetics as it did in his five normal volunteers. Six of the eight diabetic patients were also tested by intravenous injection of 25 g of glucose and showed a markedly impaired insulin secretion. Hindberg, Enk, and Persson (41) confirmed the observation that secretin induced insulin secretion is not impaired in maturity-onset diabetics. Vinik, Kalk, and Jackson (96) have indicated that the early insulin response to secretin and impure cholecyotokinin-pancreozymin (probably contaminated with GIP) Gastric Inhibiting Peptide, are normal in diabetic patients. 6. Aminophylline (Adenosine 3°5”-Monophosphate Diesterase Inhibitor Cerasi and Luft (11) have proposed that the abnormal glucose-mediated insulin response in 336 Diabetes Mellitus some prediabetics is due to a defect in the generation or cellular action of pancreatic beta cell cyclic AMP. This theory is based on studies in which they showed that aminophylline infusions given simultaneously with the glucose infusions normalized or improved the early insulin response to glucose in eight out of nine prediabetics. No such effect was noted in five patients with overt diabetes and only two of eleven nonprediabetic healthy subjects showed an increase in insulin response. Other investigators have shown that intravenous aminophylline causes many hormonal and metabolic changes in normal individuals. 7. Glucagon Simpson et al. (90) infused glucagon and 25 g of glucose over three minutes into eight maturity-onset diabetics and nine normal volunteers and measured their plasma glucose and insulin responses. Each patient had previously had a similar study done in which only glucose was given. The diabetic patients had no insulin response to glucose. The incremental insulin secretion caused by glucagon was the same in the diabetics as the normals. The incremental insulin secretion in- creased glucose utilization in the normals, but not the diabetics. 8. Serotonin Antagonists Lebovitz and Feldman (51) have hypothesized that intracellular serotinin and/or dopamine are tonic inhibitors of insulin secretion. In one of their studies (83) they attempted to determine whether the impaired glucose-mediated insulin secretion in maturity-onset diabetics was due to an exaggerated tonic effect of the proposed intracellular serotonin. They performed intravenous glucose studies on normal volunteers and maturity-onset diabetic patients. They treated both groups with a placebo and repeated the intravenous studies. Placebo treatment had no effect on glucose disposal or insulin secretion as compared to the control study. Both groups were then treated with the serotonin antagonist, methysergide maleate, and the intravenous glucose study repeated. Methysergide had no effect on insulin secretion in the volunteers, but increased insulin secretion in the diabetics by 48 percent. In a similar study (1) they have also shown that methysergide potentiates tolbutamide-mediated insulin release in diabetics (39 percent increase). SULFONYLUREA DRUGS Several properties of sulfonylurea-stimulated insulin secretion indicate that the mechanism is different from that of glucose and most other stimuli (31,36) (1) sulfonylurea drugs stimulate insulin secretion in vitro in the total absence of glucose; (2) sulfonylurea drug-mediated insulin secretion cannot be blocked in vitro or in vivo by mannoheptulose, 2 deoxy-D-glucose or diazoxide, (3) sulfonylurea drugs stimulate only the first phase of insulin release from perifused or perfused pancreas systems. Sulfonylurea drugs also potentiate glucose and amino acid mediated insulin secretion. | Malaisse et al. (61) [have suggested that sulfonylurea drugs decrease calcium efflux from beta cells and stimulate insulin secretion by increasing cytosol calcium concentrations. There is some question as to whether sulfonylurea drugs may inhibit beta cell adenosine 3”5”-mono- phosphate diesterase and act through increased beta cell cyclic AMP (86). This is somewhat un- likely as both butyrylated cyclic AMP derivatives and theophylline have an absolute requirement for glucose in order to stimulate insulin secretion (31,53). Intravenous administration of 1 g of tolbutamide to maturity-onset diabetic patients ordinarily stimulates a significantly smaller rise in plasma insulin than it does in normal individuals (23,76). Prediabetic subjects have a normal insulin secretory response to intravenous tolbutamide (3). Drugs Enhancing Insulin Secretion 337 DIRECTIONS FOR RESEARCH TO DEVELOP DRUGS THAT STIMULATE INSULIN SECRETION From the information reviewed in this chapter several obvious conclusions can be drawn. The majority of maturity-onset diabetic patients (several million in the U.S.A.) could be successfully treated if drugs were developed which either corrected the specific defect in insulin secretion which occurs in diabetes millitus or augmented nutrient stimulated insulin secretion in general, thereby allowing pharmacologic normalization of the plasma glucose. Such drugs would also be useful to prevent the development of diabetes mellitus in people with- the diabetic genetic defect. To develop these drugs it will be necessary to learn more about the mechanisms by which cells secrete granules and specifically how this process is carried out in the beta cell. More research must be done to define the specific defects in insulin secretion which occur in diabetes mellitus. Finally, we must learn how to develop drugs which alter basic processes selectively in the beta cell. Our knowledge of the mechanisms of insulin secretion are woefully inadequate. How do beta cells recognize the presence of insulinogenic stimuli? Is there a specific glucose receptor? How does glucose interact with the receptor? Do other factors influence the interaction of glucose with the receptor? How is the synthesis of the receptor controlled? If there is not a specific glucose receptor, what is the mechanism by which glucose initiates insulin secretion? Can the beta cell membrane be modified so as to increase the recognition of glucose by the beta cell? Several in vitro studies indicate that this may be possible. Lambert, Henquin, and Orci (50) have shown that preincubation of isolated islets with pronase (2 to 20 micrograms/ml) remarkably en- hance subsequent glucose-mediated insulin release. Hellman and co-workers (40) showed that a variety of sulfhydryl reagents which interact at the beta cell membrane increase both basal and glucose stimulated insulin secretion in vitro. They have also presented some data to suggest that the effects of sulfonylurea drugs on insulin secretion may occur through a similar mechanism. Even less is known about the interaction of other insulinogenic stimuli (amino acids, peptide hormones, and catecholamines) with the beta cell plasma membrane. The intracellular events which occur after the identification of the insulin secretogogue are a mystery. Is the calcium ion the final ultimate intracellular message for the granule discharge system? Techniques need to be developed to measure calcium in the cytosol. Malaisse's theory concerning the effects of hormones and cyclic AMP on intracellular calcium movement needs to be scrutinized and tested. Does cyclic AMP have an intracellular function independent of calcium ion movement? Do intracellular monoamines control insulin secretion through a tonic inhibition of granule discharge? If so, how are the intracellular monoamine levels controlled? What is the mechanism of their inhibitory action? Are there other intracellular messengers that influence insulin secretion? The function of the granule secreting system also needs to be clarified. What is the func- tion of the microtubular system? How is it controlled? How are the granules moved? Does the cell web impede the release of insulin? If so, what controls the synthesis and function of the web? In addition to basic research on the control of insulin secretion, it is necessary to carry out applied research to define the specific alterations in insulin secretion in maturity-onset diabetic patients. Glucose, amino acid, and sulfonylurea drug mediated insulin secretion are im- paired. Secretin, glucagon, and beta adrenergic agonist mediated insulin secretion seem to be normal. These data suggest that the defect in diabetes mellitus is in stimuli recognition rather 338 Diabetes Mellitus than granule release. What is the defect? Can it be modified? Do intracellular monoamines such as serotonin have anything to do with the diabetic secretory defect? Another very important area to be investigated is why agents such as glucagon, secretin, alpha adrenergic receptor antagonists, and serotonin antagonists increase glucose-mediated insulin secretion in diabetic patients, but do not improve glucose uptake. Drug development presents even greater problems to be solved. The major difficulty would appear to be to develop agents that affect fundamental cellular processes, but only in the beta cell. Can drugs be made which will affect beta cell recognition of insulinogenic stimuli without affecting other cells? For the present this question is unanswerable, but it needs to be explored. The same question can be asked about drugs that influence the cyclic AMP system. Can pancreatic beta cell cyclic AMP be preferentially increased? Smith (91), in a recent review, has discussed the evidence that drugs may preferentially inhibit the 3”5”adenosine monophosphate diesterase of one tissue. Similar data were presented with reference to activation or inhibition of adenylate cyclase. Chemicals (ionophones) which increase the flow of calcium into cells have recently been made. Their potential effects on insulin secretion are under active investigation. When more knowledge is available about the biochemistry and function of the pancreatic microtubular and micro- filamentous systems, it may be possible to develop drugs that will increase granule discharge. It is abundantly clear that there are many exciting avenues opening up for the development of drugs that will stimulate insulin secretion. Their development, however, must be based on in- creases in our fundamental knowledge about the mechanisms of insulin secretion and the defects in it that are characteristic of the diabetic state. The only useful drugs currently available that stimulate insulin secretion are the sulfo- nylureas. It is somewhat surprising that after 20 years, we still do not know how they stimulate insulin secretion, the extent of their usefulness in the treatment of patients with maturity onset diabetes mellitus, the mechanism of their chronic antidiabetic action (pancreatic versus extra-pancreatic), nor the hazards associated with their chronic use (29,34,42,81,94,95)./In spite of some suggestions that the newer generations of sulfonylurea drugs are different than the original ones (78), most studies indicate that their clinical usefulness will not be much different, and it seems unprofitable to continue developing additional analogues until answers are obtained to the questions posed above. POTENTIAL VALUE OF THIS RESEARCH The potential value of this research is immense. Approximately four million Americans have diabetes mellitus that needs treatment. Many millions more will eventually develop the disease. Oral agents that could stimulate insulin secretin and/or correct the insulin secretory defect in diabetes mellitus would revolutionize the treatment. One could hope to be able to more easily normalize the blood sugar or even perhaps prevent the development of the disease. The effect that this would have in minimizing the complications and sequelae of the disease have been covered in other sections of this monograph. The basic information on the control of the secretory process would be useful in many other areas of medicine and biomedical research. Drugs Enhancing Insulin Secretion 339 REFERENCES 1. Baldridge, JA, KE Quickel, Jr, JM Feldman, and HE Lebovitz 1974. Potentiation of tolbutamide- mediated insulin release in adult onset diabetes by methysergide maleate. Diabetes 23:21-24. 2. Bodanszky, M 1974. Gastrointestinal hormones, families of oligeoelectrolytes. Endocrinology of the Gut. Chey and Brooks (eds) Thorofare, NJ Charles and B Slack. pp 3-13. 3. Boden, G, JS Soeldner, RE Gleason, and A Marble 1968. Elevated serum growth hormone and de- creased insulin in prediabetic males after intravenous glucose and tolbutamide. J Clin Invest 47:729-739. 4, Brisson, GR, and WJ Malaisse 1973. The stimulus secretion coupling of glucose-induced insulin release. XI Effects of theophylline and epinephrine on “3Ca efflux from perifused islets. Metabolism 22:455-465. 5. Brisson, GR, F Malaisse-Lagae, and WJ Malaisse 1972. The stimulus secretion coupling of glucose-induced insulin release. VII. A proposed site of action for adenosine 3°5”-cyclic monophosphate. J Clin Invest 51:232-241. 6. Brown, JC, JR Dryburgh, and RA Pederson 1974. Gastric inhibitory polypeptide (GIP). Endo- crinology of the Gut. Chey and Brooks (eds) Thorofare, NJ Charles and B Slack. pp 76-81. 7. Buse, MG, AH Johnson, D Kuperminc, and J Buse 1970. Effect of a-adrenergic blockade on insulin secretion in man. Metabolism 19:219-225. 8. Cegrell, L 1968. The occurrence of biogenic monoamines in the mammalian endocrine pancreas. Acta Physiol Scand Suppl 314:1-60. 9. Cerasi, E, and R Luft 1967a. The plasma insulin response to glucose infusion in healthy sub- jects and in diabetes mellitus. Acta Endocrinol 55:278-304. 10. Cerase, E, and R Luft 1967b. Insulin response to glucose infusion in diabetic and nondiabetic monozygotic twin pairs. Genetic control of insulin response? Acta Endocrinol 55:330-345. 11. Cerasi, E, and R Luft 1969. The effect of an adenosine 3”5”-monophosphate diesterase in- hibitor (aminophylline) on the insulin response to glucose infusion in prediabetic and diabetic subjects. Horm Metab Res 1:162-168. 12. Cerasi, E, and R Luft 1970. The pathogenesis of diabetes mellitus--a proposed concept. Pathogenesis of Diabetes Mellitus. 13th Nobel Symposium. Cerasi and Luft (eds) Almqvist and Wiksell, Stockholm. pp 17-39. 13. Cerasi, E, and R Luft 1974. Diabetes mellitus--a disease of pancreatic and extrapancreatic origin. Adv Metab Disord 7:193-210. 14. Cerasi, E, S Efendic, and R Luft 1969. Role of adrenergic receptors in glucose-induced insulin secretion in man. Lancet 2:301-302. 15. Charles, MA, R Fanska, FG Schmid, PH Forsham, and GM Grodsky 1973. Adenosine-3"5"-mono- phosphate in pancreatic islets: glucose induced insulin release. Science 179:569-571. 16. Coore, HG, and PJ Randle 1964a. Inhibition of glucose phosphorylation by mannoheptulose. Biochem J 91:56-59. 17. Coore, HG, and PJ Randle 1964b. Regulation of insulin secretion studied with pieces of rabbit pancreas incubated in vitro. Biochem J 93:66-78. 18. Coore, HG, PJ Randle, E Simon, PF Kraicer, and MC Shelesnyak 1963. Block of insulin secretion by D-mannoheptulose. Nature 197:1264-1266. 19. Dean, PM, and EK Matthews 1968. Electrical activity in pancreatic islet cells. Nature 219: 389-390. 340 Diabetes Mellitus 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34, 35. 36. 37. 38. x9. Dean, PM, and EK Matthews 1970a. Glucose-induced electrical activity in pancreatic islet cells. J Physiol (London) 210:255-264. Dean, PM, and EK Matthews 1970b. Electrical activity in pancreatic islets: effect of ions. J Physiol (London) 210:265-275. Deckert T 1968. Insulin secretion following administration of secretin in patients with diabetes mellitus. Acta Endocrinol 59:150-158. Deckert, T, and P Mogensen 1970. Plasma insulin after tolbutamide in diabetes and non- diabetes. Acta Med Scand 187:309-312. Deckert, T, UB Lauridsen, SN Madsen, and M Deckert 1972. Serum insulin following iso- prenaline in normal and diabetic persons. Horm Metab Res 4:229-232. Efendic, S, E Cerasi, and R Luft 1973. Effect of blockade of the a-adrenergic receptors on insulin response to glucose infusions in prediabetic subjects. Acta Endocrinol 74:542-547, Fajans, SS, and JC Floyd, Jr 1972. Stimulation of islet cell secretion by nutrients and by gastrointestinal hormones released during digestion. Handbook of Physiology, Section 7, Endocrinology Vol 1 Endocrine Pancreas, Washington, Amer Physiol Soc. pp 473-493. Fajans, SS, JC Floyd, Jr, JW Conn, S Pek, J Rull, and RF Knopf 1969. Plasma insulin responses to ingested glucose and to infused amino acids in subclinical diabetes and prediabetes. Diabetes. Proc VIth International Congress Diabetes Federation. J Ostman and RDG Milner (eds) Excerpta Medica, Amsterdam, pp 515-522. Fajans, SS, R Quibrera, S Pek, JC Floyd, Jr, HN Christensen, and JW Conn 1971. Stimulation of insulin release in the dog by a nonmetabolizable amino acid. Comparison with leucine and arginine. J Clin Endocrinol Metab 33:35-41. Feldman, JM, and HE Lebovitz 1969. Appraisal of the extrapancreatic actions of sulfonylureas. Arch Intern Med 123:314-322. Feldman, JM, and HE Lebovitz 1970. Mechanism of epinephrine and serotonin inhibition of insulin release in the golden hamster in vitro. Diabetes 19:480-486. Feldman, JM, and HE Lebovitz 1971. Ammonium ion, a modulator of insulin secretion. Amer J Physiol 221:1027-1032. Feldman, JM, AE Boyd III, and HE Lebovitz 1971. Structural determinants of catecholamine action on in vitro insulin release. J Pharmacol Exp Ther 176:611-621. Feldman, JM, KE Quickel Jr, and HE Lebovitz 1972. Potentiation of insulin secretion in vitro by serotonin antagonists. Diabetes 21:779-788. Feldman, R, D Crawford, R Elashoff, and A Glass 1974. Oral hypoglycemic drug prophylaxis in asymptomatic diabetes. Diabetes. Proc VIII Cong I.D.F. WJ Malaisse, J Pirart, and J Vallance- Owen (eds) Amsterdam, Experpta Medica, pp 574-587. Floyd, JC, Jr, SS Fajans, JW Conn, C Thiffault, RF Knopf, and E Guntsche 1968. Secretion of insulin induced by amino acids and glucose in diabetes mellitus. J Clin Endocrinol Metab 28:266-276. Grodsky, GM 1970. Insulin and the pancreas. Vitamins and hormones, New York Academic Press, pp 37-101. Grodsky, GM, and LL Bennett 1966. Cationic requirements for insulin secretion in the isolated perfused pancreas. Diabetes 15:910-913. Grodsky, GM, AA Batts, LL Bennett, C Voella, NB McWilliams, and DF Smith 1963. Effects of carbohydrates on secretion of insulin from isolated rat pancreas. Amer J Physiol 205:638-644. Grodsky, GM, JH Karam, FC Pavlatos, and PH Forsham 1965. Serum insulin response to glucose in prediabetic subjects. Lancet 1:290. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. Drugs Enhancing Insulin Secretion 341 Hellman, B, L Idahl, A Lernmark, J Sehlin, and I Tdljedal 1974. Membrane sulphydryl groups and the pancreatic beta cell recognition of insulin secretagogues. Diabetes. Proc VIII Cong. I.D.F., WJ Malaisse, J Pirat, and J Vallance-Owen (eds) Amsterdamn, Excerpta Medica, pp 65-78. Hindberg, I, B Enk, and I Persson 1970. Insulin stimulation by secretin in diabetes. Effects of repeated and varied doses. Horm Metab Res 2:131-134. Keen, H, RJ Jarrett, and JH Fuller 1974. Tolbutamide and arterial disease in borderline diabetes. Diabetes. Proc VIII Cong I.D.F., WJ Malaisse, J Pirart, and J Vallance-Owens (eds) Amsterdam, Excerpta Medica, pp 588-602. Kipnis, DM 1969. Insulin antagonism and diabetes mellitus. Diabetes. Proc VI Cong I.D.F., Amsterdam, Excerpta Medica, pp 257-276. Kuo, WN, DS Hodgins, and JF Kuo 1973. Adenylate cyclase in islets of Langerhans. Isolation of islets and regulation of adenylate cyclase activity by various hormones and agents. J Biol Chem 248:2705-2711. Lacy, PE, 1967. The pancreatic beta cell. New Engl J Med 276:187-194. Lacy, PE, 1970. Beta cell secretion from the standpoint of a pathobiologist. Diatetes 19: 895-905. Lacy, PE, and WJ Malaisse 1973. Microtubules and beta cell secretion. Recent Prog Horm Res 29:199-221. Lacy, PE, MM Walker, and CJ Fink 1972. Perifusion of isolated rat islets in vitro. Partici- pation of the microtubular system in the biphasic release of insulin. Diabetes 21:987-998. Lacy, PE, NJ Klein, and CJ Fink 1973. Effect of cytochalasin B on the biphasic release of insulin in perifused rat islets. Endocrinology 92:1458-1468. Lambert, AE, JC Henquin, and L Orci 1974. Role of beta cell membrane in insulin secretion. Diabetes. Proc VIII Cong I.D.F., WJ Malaisse, J Pirart, and J Vallance-Owen (eds) Amsterdam, Excerpta Medica pp 79-94. Lebovitz, HE, and JM Feldman 1973. Pancreatic biogenic amines and insulin secretion in health and disease. Fed Proc 32:1797-1802. Loubatiéres-Mariani, MM, J Chapal, R Alfric, and A Loubati&res 1973. Studies of the cholinergic receptors involved in the secretion of insulin using isolated perfused rat pancreas. Dia- betologica 9:439-446. Malaisse, WJ 1972. Hormonal and environmental modification of islet activity. Handbook of Physiology Section 7, Endocrinology, Vol 1. Endocrine Pancreas. Washington, American Physio- logocal Society, pp 237-260. Malaisse, WJ 1973. Insulin secretion: multifactorial regulation for a single process of release. Diabetologica 9:167-173. Malaisse, WJ, F Malaisse-Lagae, and D Mayhew 1967a. A possible role for the adenylcyclase sys- tem in insulin secretion. J Clin Invest 46:1724-1734. Malaisse, W, F Malaisse-Lagae, P Wright, and J Ashmore 1967b. Effects of adrenergic and cholinergic agents upon insulin secretion in vitro. Endocrinology 80:975-978. Malaisse, WJ, F Malaisse-Lagae, and G Brisson 1971. The stimulus secretion coupling of glucose-induced insulin release. II. Interaction of alkali and alkaline earth cations. Horm Metab Res 3:65-70. Malaisse, WJ, M Mahy, GR Bresson, and F Malaisse-Lagae 1972a. The stimulus secretion coupling of glucose-induced insulin release. VIII. Combined effects of glucose and sulfonylureas. Europ J Clin Invest 2:85-90. 342 Diabetes Mellitus 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 7Z. 78. Malaisse, WJ, DL Hager, and L Orci 1972b. The stimulus-secretion coupling of glucose-induced insulin release. IX Participation of the beta cell webb. Diabetes 21: (suppl 2:) 595-604. Malaisse, WJ, GR Brisson, and LE Baird 1973a. Stimulus secretion coupling of glucose-induced insulin release. X. Effect of glucose on *°Ca efflux from perifused islets. Amer J Physiol 224:389-394. Malaisse, WJ, DG Pipeleers, and M Mahy 1973b. The stimulus secretion coupling of glucose- induced insulin release. XII. Effects of diazoxide and gliclazide upon "°Ca efflux from perifused islets. Diabetologica 9:1-5. Malaisse-Lagae, F, GR Brisson, and WJ Malaisse 1971. The stimulus secretion coupling of glucose-induced insulin release. VI. Analogy between the insulinotropic mechansims of sugars and amino acids. Horm Metab Res 3:374-378. Malins, JM, 1972. Diabetes in the population. Clinics in Endocrinology and Metabolism 1:645-672. Marble A, 1971. Current concepts of diabetes. Joslin's Diabetes Mellitus 11th edition. Philadelphia, Lea and Febiger, pp 1-19. Marks, HH, LP Krall, and P White 1971. Epidemiology and detection of diabetes. Joslin's Diabetes Mellitus 11th edition. Philadelphia, Lea and Febiger, pp 10-34. Matschinsky, FM, R Landgraf, J Ellerman, and J Kotler-Brajtburg 1972. Glucoreceptor mech- anisms in islets of langerhans. Diabetes 21 (suppl 2):555-569. Matschinsky, FM, and J. Ellerman 1973. Dissociation of the insulin releasing and the metabolic functions of hexoses in islets of langerhans. Biochem Biophys Res Comm 50:193-199. McDaniel, ML, S King, S Anderson, C Fink, and PE Lacy 1974. Effect of cytochalasin B on hexose transport and glucose metabolism in pancreatic islets. Diabetologica 10:303-308. Milner, RDG 1969. Stimulation of insulin secretion in vitro by essential amino acids. Lancet 1:1075. Milner, RDG, and CN Hales 1967. The role of calcium and magnesium in insulin secretion from rabbit pancreas studied in vitro. Diabetologica 3:47-49. Misbin, RI, PJ Edgar, and DH Lockwood 1970. Adrenergic regulation of insulin secretion during fasting in normal subjects. Diabetes 19:688-693. Orci, L, KH Gabbay, and WJ Malaisse 1972. Pancreatic beta cell web: its possible role in in- sulin secretion. Science 175:1128-1130. Orci, L, M Ravazzola, M Amherdt, and F Malaisse-Lagae 1974. The beta cell boundary. Diabetes. Proc. VIII Cong. I.D.F. WJ Malaisse, J Pirart and J Vallance-Owen (eds) Amsterdam, Excerpta Medica, pp 104-118. Parker, ML, RS Pildes, K Chao, M Cornblath, and DM Kipnis 1968. Juvenile diabetes mellitus, a deficiency in insulin. Diabetes 17:27-32. Perly, M, and DM Kipnis 1967. Plasma insulin responses to oral and intravenous glucose: studies in normal and diabetic subjects. J Clin Invest 46:1954:1962. Perly, M, and DM Kipnis 1966. Plasma insulin response to glucose and tolbutamide of normal weight and obese diabetic and non-diabetic subjects. Diabetes 15:867-874. Persson, I, and L Larsen 1972. Serum insulin concentration after alpha adrenergic blockade and secretin in diabetes. Acta Endocrinol 71:331-337. Pfeiffer, EF, 1974. Sulfonlyureas: newer aspects of pharmacology and clinical efficacy. Diabetes. Proc VIII Cong I.D.F., WJ Malaisse, J Pirart, and J Vallance-Owen (eds) Amsterdam, Excerpta Medica, pp 563-573. 79. 80. 81. 82. 83. 84. 85. 86. 37. 88. 89. 90. 91. 92. 93. 94. 95. 96. 7. 98. Drugs Enhancing Insulin Secretion 343 Porte, D, Jr, 1967. Beta adrenergic stimulation of insulin release. Diabetes 16:150-155. Porte, D, Jr, and RP Robertson 1973. Control of insulin secretion by catecholamines, stress and the sympathetic nervous system. Fed Proc 32:1792-1796. Prout, TE, 1974. Adverse effects of oral hypoglycemic drugs. Diabetes. Proc VIII I.D.F., WJ Malaisse, J Pirart, and J Vallance-Owen (eds) Amsterdam, Excerpta Medica, pp 612-623. Pyke, DA, and KW Taylor 1967. Glucose tolerance and serum insulin in unaffected identical twins of diabetes. Brit Med J 4:21-22. Quickel, KE, Jr, JM Feldman, and HE Lebovitz 1971. Enhancement of insulin secretion in adult onset diabetes by methysergide maleate; evidence for an endogenous biogenic monoamine mechanism as a factor in the impaired insulin secretion in diabetes mellitus. J Clin Endo- crinol Metab 33:877-881. Randle, PJ, and CN Hales 1972. Insulin release mechanisms. Handbook of Physiology, Section 7 Endocrinology, Vol 1 Endocrine Pancreas. Washington, American Physiologic Society, pp 219- 235. Robertson, RP, and D Porte, Jr, 1973. The glucose receptor. A defective mechanism in dia- betes mellitus distinct from the beta adrenergic receptor. J Clin Invest 52:870-876. Roth, J, 1971. Sulfonlyureas: effects in vivo and in vitro. Ann Intern Med 75:607-621. Sanger, JW, and H Holtzer 1972. Cytochalasin B: effects on cell morphology, cell adhesion and mucopolysaccharide synthesis. Proc Nat Acad Sci (USA) 69:253-257. Seltzer, HS, EW Allen, AL Herron, Jr, and MT Brennan 1967. Insulin secretion in response to glycemic stimulus. Relation of delayed initial release to carbohydrate intolerance in mild diabetes mellitus. J Clin Invest 46:323-335. Sharp, R, S Culbert, J Cook, A Jennings, and IM Burr 1974. Cholinergic modification of glucose-induced biphasic insulin release in vitro. J Clin Invest 53:710-716. Simpson, RG, A Benedetti, GM Grodsky, JH Karam, and PH Forsham 1966. Stimulation of insulin release by glucagon in noninsulin-dependent diabetics. Metabolism 15:1046-1049. Smith, CG, 1974. The cyclic AMP system and drug development. Advances in Enzyme Regulation, G Weber (ed). In press. Soeldner, JS, RE Gleason, L Rojas, CB Kahn, and A Marble 1969. Serum insulin and serum insulin-blood glucose relationship in genetic prediabetic males with normal glucose tolerance. Diabetes. Proc VI Cong I.D.F., J Ostman and RDG Milner (eds) Amsterdam, Excerpta Medica Foundation, pp 504-514. Turtle, JR, and DM Kipnis 1967. An adrenergic receptor mechanism for the control of cyclic 3”5”adenosine monophosphate synthesis in tissues. Biochem Biophys Res Comm 28:797-802. University Group Diabetes Program 1970. I--Design, methods and baseline characteristics. II--Mortality results. Diabetes 19 (suppl. 2):747-830. University Group Diabetes Program 1971. Diabetes III Clinical implication of the UGPD results. J Amer Med Assoc 218:1400-1410. Vinik, AI, WJ Kalk, and WPU Jackson 1974. A unifying hypothesis for hereditary and acquired diabetes. Lancet 1:485-486. Wrenshall, GA, A Bogoch, and RC Ritchie 1952. Extractable insulin of pancreas. Diabetes 1:87-107. Yalow, RS, and SA Berson 1960. Plasma insulin concentration in nondiabetic and early diabetic subjects. Diabetes 9:254-260. 25 "DRUGS ALTERING CARBOHYDRATE AND LIPID METABOLISM Bernard Robert Landau INTRODUCTION With increasing delineation of the biochemical alterations that occur in diabetes mellitus, new pharmacological approaches to its therapy become possible. These approaches differ in the available data supporting their rationale, the biological systems for testing a potential agent, and the "leads" and agents already at hand. In Table 1 many of these leads and agents are grouped according to their presumed mechanisms of action. While the approaches that follow are referenced in terms of the biochemical process to which the pharmacological agent is or would be directed, alterations in one process invariably re- sult in the alterations in other processes, since these are integrated. Many of the pharmacologi- cal agents now under study, while presumed to act on one process, appear to act at several loci, and the primary site of action for most leads or even established agents is still uncertain. The agents used clinically and other leads now available are inhibitors. The approaches pro- posed are therefore generally directed toward inhibition rather than stimulation of processes. Inhibitors of biochemical processes are much easier to develop than stimulators, presumably because of the high specificity usually encountered in the activation of biological processes. Progress toward agents which, for example, mimic insulin action on membranes or directly alter an enzyme to increase its activity, will depend upon further definition at the molecular level of these processes and the ingenuity of the chemist using this information in synthesizing compounds that reproduce their function. Further, agents may have to be developed selective in their action toward a given tissue. Despite the knowledge we already have of mammalian membrane structure, transport processes, and hormone action, our knowledge is still woefully inadequate for the design of such agents. Perhaps more important for any approach toward the prevention, retarding, or reversal of the chronic complications of diabetes, and at least one of these should be the major objective of any approach if it is to provide a major addition to the therapy of diabetes mellitus, is the need for models by which potential agents can be tested for their effectiveness. While measurements in man of basement membrane thickening, conjunctival vessel alterations, etc., may be employed, animal models are needed which allow measurement of responses to agents over shorter. durations than in man. Toward this end, lesions similar to those seen in human nephropathy and retinopathy have been reported in animals, but the utility of these preparations is yet to be established. Most important, the development through basic laboratory procedures of potential agents for the prevention and/or therapy of diabetes mellitus may not require the largest fraction of the in- vestment in manpower and financial resources. As is illustrated by the University Group Diabetes Program, the establishment of efficacy for a drug will almost certainly require a very large in- vestment over many years. Therefore, the finding of ''leads' may not prove to be the largest obstacle to success, but rather it may be whether an individual pharmaceutical company or any other 344 Presumed Mechanism of Action Inhibiting Glucose Absorption TABLE 1. Therapy of Diabetes Mellitus Compound (s) Phenformin Quinolinic Acid References Kruger et al., 1970; Caspary and Creutzfeldt, 1973 Veneziale et al., 1967 wm 3 Biguanides Haeckel and Haeckel, 1972; Altschuld and Kruger, 1968 Inhibiting Gluconeogenesis Pent-4-enoic Acid Ruderman et al., 1970, Toews et al., 1970 Metyrapone Henke and Doe, 1967 Trypotophan Mirsky et al., 1957 Enhancing Insulin Action Indole Acetic Acid Mirsky and Diengott, 1956 Biguanides Davidoff, 1973 Inhibiting Glucagon Release Diphenylhydantoin Gerich et al., 1972 Inhibiting Glucagon Action Inhibiting Growth Hormone Release Inhibiting Respiration Inhibiting Lipolysis Inhibiting Fatty Acid Oxidation Inhibiting Aldose Reductase Inhibiting Cholesterol Synthesis Unknown Des-Histidine Glucagon A Polypeptide Biguanides Dinitrophenol Salicylates Clofibrate Nicotinic Acid Propanolol Dimethylpyrazole Pent-4-enoic Acid (+) -Decanoylcarnitine Diphenylene iodonium Dichloroacetic Acid Glutaric Acid Derivatives Clofibrate Pyridinium Chlorides Y-guanidinobutyramide (HL 523) Lande et al., 1972 Brazeau et al., 1973 Williams et al., 1967; Davidoff, 1968 Steward and Hanley, 1969 Steward and Hanley, 1969 Hoppel, 1973 Carlson, 1969 Hoppel, 1973 Gerritsen and Dulin, 1965; Hollobaugh et al., 1967 Toews et al., 1970; Corredor et al., 1968 Williamson et al., 1968; Williamson et al., 1969; McGarry and Foster, 1973 Steward and Hanley, 1969 Stacpoole and Felts, 1971 Gabbay, 1973; Morrison and Winegard, 1973; Gabbay and Kinoshito, 1972 Bergquist, 1970; Harrold et al., 1969 Fanshawe et al., 1970; Blickens and Riggi, 1969 Butterfield et al., 1969; Schless et al., 1970 ¢pe wsTTOqER3OW PTAIT pu® ayexphyoqaey burrelTV Sbnia 346 Diabetes Mellitus similar sized group has the capacity to perform the required clinical testing and, if not, whether other long-term support can be established so that the finding of new pharmacological agents of real benefit can be brought to fruition. APPROACHES The approaches that follow (Fig. 1) can be classified in terms of decreasing glucose production or increasing glucose utilization, either one of which will lower blood glucose concen- tration. In the former group would be the approach of (I) decreasing the absorption of glucose, and (II) decreasing glucose formation (gluconeogenesis) either by inhibiting the catalysts (en- zymes) required for formation, or decreasing the availablity of precursors (substrates) needed to form the glucose. In the latter group would be the development of an agent (III) mimicking the acting of insulin, (IV) preventing or retarding the destruction of insulin when present, (V) pre- venting hormone actions that are counter to insulin's action, (VI) decreasing the yield of energy from glucose metabolism (respiration) so that more glucose would have to be utilized to provide an equivalent quantity of energy, and (VII) preventing the utilization of fat by inhibiting the breakdown of fat (lipolysis) or (VIII) its metabolism (fatty acid oxidation) so that increased quantities of glucose would have to be used as a substitute for the fat and (IX) decreasing blood glucose and/or fatty acid concentration through regulatory controls in the central nervous system. Inhibition of specific pathways of glucose utilization, that is inhibition of (X) the polyol pathway and (XI) glycoprotein formation, and (XII) decreasing cholesterol and triglyceride forma- tion are approaches directed toward preventing the formation of components believed to contribute to the long-term complications seen in the diabetic. A broad screening approach (XIII) would encompass all of the above approaches and others not recognized. I. GLUCOSE ABSORPTION Rationale: If glucose absorption is decreased, blood glucose concentration will rise less on carbohydrate ingestion. Discussion: The diabetic should ingest a normal quantity of carbohydrate, since there is evidence that when a low quantity of carbohydrate is ingested, glucose tolerance, that is the ability to utilize glucose, is impaired. Further, with the inhibition of absorption of any signifi- cant quantity of carbohydrate, diarrhea should ensue. This approach then has very little, if any, promise. Systems: A number of procedures for measuring glucose transport in vitro (intestinal sacs, strips, etc.) and in vivo (intestinal intubations) are well established. Leads: Phenformin (B1Y has been shown to inhibit glucose absorption in man (18,56). Since intestinal glucose transport is energy dependent, decreased absorption may be consequent to the effect of phenformin on the respiration (see Approach VI) of the intestinal epithelium. In accord with this, intestinal amino acid transport is also inhibited (18). The effect on intestinal transport is unlikely to contribute significantly to phenformin's overall hypoglycemic effect. IT. GLUCONEOGENESIS Rationale: Since the diabetic's elevated blood glucose concentration is due to overproduction as well as underutilization of glucose, decrease in glucose formation should lower blood glucose concentration. Glucosuria should then be diminished and, if the chronic complications of diabetes are due to elevated blood glucose concentrations, these should be lessened. Since the Drugs Altering Carbohydrate and Lipid Metabolism 347 CHOLESTEROL il TRIGLYCERIDE x, ?CNS FACTORS LIVER XII. Vill. L ( Plog | receiving |0'5¢ Oo Oo 13 GLUCOSE |. ABSORPTION FIGURE 1. Possible sites of action (I - XIII) of drugs altering carbohydrate and lipid metabolism. 348 Diabetes Mellitus liver is the prime site of glucose formation (the kidney and possibly intestine participating under some circumstances), an inhibitor of hepatic glucose formation should be sought. Discussion: Since the defect in the diabetic is also decreased glucose utilization, this approach in general is not attractive for the diabetic. An elevated blood glucose concentration, at least at concentrations near 300 mg percent or 400 mg percent, favors through mass action glucose utilization at a similar rate to that at 100 mg percent in the presence of insulin (78). While glucose utilization at these elevated concentrations will be directed relatively more in the diabetic than in the normal to insulin-independent tissues, a lowering of blood glucose con- centration solely by decreasing production of glucose in the diabetic should result in a still further decrease in glucose utilized by the peripheral tissues, muscle, and fat. Thus, increased mobilization of fatty acids, inhibition by fatty acids of glucose utilization by muscle, and in- creased ketone formation by liver would be expected. These expectations are perhaps realized in alcoholics who present themselves in ketoacidosis, but with relatively low blood glucose concentra- tion, ethanol presumably having decreased gluconeogenesis. Administration of glucose and fluids is often the only required treatment. Inhibition of gluconeogenesis may prove of some therapeutic value in such acute emergencies as the occasional hyperosmolar coma, where glucose concentrations of 800 mg percent or more are encountered, but insulin and fluids seem to offer for these circum- stances a much better therapeutic approach. A decrease in blood glucose concentration could attenutate some of the load on the beta cells of the pancreas and could reduce the diversion of glucose to such pathways as the polyol pathway (see Approach X). Systems: Key enzymes in the control of gluconeogenesis have been identified, isolated, and purified so that screening for agents that inhibit them is possible. Much is already known of the processes regulating gluconeogenesis so that alterations in the activities of concentrations of natural factors which stimulate or inhibit gluconeogenesis should be possible. Liver (and kidney) slices and perfusion systems have been refined and are available for screening for, and testing of, potential agents, and there are procedures for measuring glucose production in animals and man both directly and using isotopes as tracers. Leads: Ethanol administration decreases blood glucose concentration presumably by decreasing the ‘availability of substrate for gluconeogenesis. Its ability to decrease glucose production has been shown in liver slices and its ability to lower blood glucose concentration has been shown in man (47). Quinolinic acid inhibits phosphoenolpyruvate carboxykinase, a key enzyme required for gluconeogenesis, and in accord with this it inhibits glucose formation from pyruvate by liver (87). Phenformin, along with its other possible modes of actions (see Approaches I, IV, and VI), has been shown to inhibit gluconeogenesis in perfused liver of the rat and guinea pig (1,45). Phenformin is the phenylethyl derivative of biguanide and is prescribed in the United States as an oral hypoglycemic agent for the management of the diabetic. Other derivatives of biguanides "are prescribed in other countries. They are similar in their actions, although potencies vary. In man phenformin increases glucose utilization. This appears to be associated with an in- crease rather than a decrease in glucose production (55,60,61). However, it is postulated that the increase in glucose production is less than would normally be expected in response to the increased utilization, and thus phenformin's inhibition of gluconeogenesis contributes to its hypoglycemic action. Drugs Altering Carbohydrate and Lipid Metabolism 349 Pent-4-enoic acid also inhibits hepatic gluconeogenesis (73,85), but this appears to be through its primary action of inhibiting fatty acid oxidation (see Approach VIII). Antagonists of glucocorticoids would be expected to inhibit gluconeogenesis, but the likely response of the pituitary then to increase glucocorticoid production and the dangers of inadver- tently producing adrenal insufficiency with an excessive dose, if effective, would probably severly limit such an agent. Metyrapone, which reduces cortisol production by inhibition of adrenal 11- B-hydroxylation, could serve as a lead (49). ITI. INSULIN MIMICS Rationale: Since diabetes mellitus is a disease of relative insulin deficiency, an agent which had all the actions of insulin and could be taken orally would provide effective and con- venient therapy. Discussion: The extent of the contribution such an agent would make depends upon whether insulin deficiency occurs concomitant with the occurrence of diabetic vascular complication or is responsible for the complications. An oral agent that mimics insulin's action would be pre- sented first to the liver, as is the case for insulin released from the pancreas, rather than to peripheral tissues as occurs with exogenous insulin administration. The difference between parenteral and oral administration may be significant in the control of blood glucose concentra- tion and in the prevention of complications. As a minimum such an oral agent would offer a con- venience not possible with parenteral administration. Systems: Preparations containing insulin receptors have now been isolated (37,77). Some information on the structure of the membrane in relation to the receptor are at hand. There are many systems in vivo and in vitro for characterizing the actions of insulin. Leads: Insulin itself, and various preparations of insulin (36) have been given orally but, while there is some absorption, their use appears impractical because of extensive destruction in, and very limited absorption of, polypeptides by the intestinal tract. IV. ENHANCING INSULIN'S ACTION Rationale: An alternative approach to the enhancing of insulin secretion and synthesis in the maturity onset diabetic would be an agent that increases the effect of whatever quantity of insulin is present. One such approach would be through the inhibition of insulin degradation. Since the largest fraction of insulin degradation appears to be in the liver, with the largest portion of the remainder occurring in the kidney and intestine, agents could be sought which inhibit degradation at these sites. Alternatively, agents could be sought which increase the effect of insulin at its site of action. Discussion: The maximum effect of such an agent is limited by the maximum quantity of insulin being secreted by the pancreas. Thus, such an agent should not be effective in the juvenile dia- betic, Systems: A protease with reasonable specificity for insulin was demonstrated by Mirsky et al. (65). A glutathione transhydrogenase in liver has been shown to cleave the A and B chains of insulin and is believed to participate in the initiation of insulin breakdown, but the enzyme may also catalyze other reactions (52). Recently proteases, claimed to be relatively specific for insulin degradation (12,13), although glucagon is also degraded (27), have been reported to be present in muscle as well as in kidney and liver. Techniques for measuring insulin by 350 Diabetes Mellitus immunoassay and its degradation using iodine tracers are available. The effects of insulin on many isolated tissue preparations, as well as intact animals, have been demonstrated, and agents could be screened in these systems to see if they enhance the effects. Leads: Tryptophan inhibits insulinase and reduces glucose concentrations in normal but not in the alloxan-diabetic rat (65). Other compounds containing the indole ring also have produced hypoglycemia in animals (64). Biguanides in low concentrations have been reported to augment or amplify insulin action. They have been hypothesized to do this through an action on metal binding sites on membranes (24) (and plasma membrane preparation may be used for initial screening of agents acting in this manner). V. INHIBITION OF COUNTER REGULATORY HORMONE SECRETION AND ACTION Rationale: Several hormones are known to counter insulin action, and therefore there would be enhanced insulin action if any of these were inhibited. Evidence has accrued that there may be exaggerated effects of counterregulating hormones in diabetics. Discussion: Two hormones appear from present information to provide the most suitable targets. Glucagon has been hypothesized to play an essential role in the pathogenesis of dia- betes mellitus (86). There is a decreased suppression of glucagon by glucose in the diabetic, relative or absolute hypergluconemia has been identified in every form of endogenous hyperglycemia, and insulin lack when glucon is suppressed does not cause endogenous hyperglycemia nor ketoacidosis (39,41). Inhibition of glucagon's release would then be expected to ameliorate or prevent the disease. A similar statement can be made for growth hormone which has been implicated in the onset of ketosis in the diabetic, has been shown to respond excessively to exercise in the diabetic, and has been proposed to play a role in the development of diabetic retinopathy. Glucocorticoid antagonists are considered in Approach II. Systems: Glucagon receptor preparations are now being developed which could allow screening for agents in vitro. Immunoassays for growth hormone and glucagon exist. Leads: Diphenylhydantoin (Dilantin®) inhibits glucagon secretion in vitro, and also inhibits insulin secretion (38). In maturity onset diabetics, diphenylhydantoin can exacerbate hyper- glycemia, but in juvenile diabetics, since the beta cell is inoperative, diphenylhydantoin might be beneficial. Derivatives of glucagon are also shown which inhibit its action (57). L-Dopa stimulates growth hormone (9) so that analogs inhibiting this action might be sought. Most attractive as a lead at present is a polypeptide from ovine hypothalamus that inhibits growth hormone secretion (10) and is called somatostatin (11). It also inhibits the release of insulin and glucagon (54). Because of the roles hypothesized for glucagon and growth hormone in the diabetic syndrome, clinical studies have begun. It was in using somatostatin that evidence was obtained that glucagon is essential for the development of ketoacidosis (41). Its effect on ketoacidosis may be at least in part through direct inhibition of hepatic ketogenesis rather than via glucagon suppression (40). Somatostatin has also been reported to suppress growth hormone levels in acromegalies (93) and insulin hypersecretion in a patient with pancreatic islet cell carcinoma (22). Evidence has been obtained for the presence of somatostatin in human as well as rat brain and pancreas (68,69). Since somatostatin is a naturally occurring polypeptide, it might be expected to have less toxicity than chemicals foreign to the human body. Toxic effects of somatostatin on platelet ag- gregation appear to be of concern at concentrations above those being used in the clinical Drugs Altering Carbohydrate and Lipid Metabolism 351 studies (28). While somatostatin has been said to perhaps be "twice blessed" by inhibiting both growth hormone and glucagon hormone levels, its half life in blood is less than four minutes, so that long-term clinical trials directed in particular toward the therapy microangiopathy await the preparation of a suitable long-acting derivative (15). In addition, analogs of somastatin which are specific for causing inhibition of glucagon release and inhibition of growth hormone release need to be synthesized and tested. The gestagen, medroxyprogesterone acetate, has been reported through its action upon pituitary function to diminish and confine the progressive course of diabetic retinopathy (16). RESPIRATION Rationale: Since the oxidation of glucose to carbon dioxide and water yields about 18 times as much useful energy in the form of ATP as its metabolism to lactate, if respiration (the utilization of oxygen with the production of ATP) is reduced, the body must utilize more glucose to produce similar quantities of energy. Discussion: The process to be inhibited is critical process for maintenance of body function, and an agent affecting the process would be expected to have dangerous side reactions. For the success of such an approach one might have to be selective in the tissues affected, since certain tissues may be harmed by not being able to increase the utilization of glucose adequately in response to the inhibition. Systems: Systems in vitro (mitochondria) and in vivo for measuring effects on respiratory function are available. Leads: Phenformin inhibits respiration in vitro and has been postulated to produce its hypo- glycemic effect in vivo by this mechanism (23,82.88). While this effect, as noted above, on theoretical grounds might be considered dangerous, phenformin has proved to be remarkably safe (without including here a consideration of the results of the University Group Diabetes Program Study). Lactic acidosis may be exacerbated by phenformin administration. " Dinitrophenol uncouples respiratory function; that is, oxygen consumption continues in its presence but with a lesser yield of ATP. Salicylates, as aspirin, also uncouple oxidative phos- phorylation. Both these compounds lower blood glucose concentrations in animals and in man under selected conditions, but of themselves they do not have any potential as hypoglycemic agents (82). There is a suggestion that salicylates reduce the incidence of diabetic retinopathy (70), but this could be associated with their effect on platelet adhesion (48,67). Clofibrate, a cholesterol and triglyceride lowering agent, recently has also been shown to inhibit respiration (51). y1l. FLIPOLYSIS Rationale: Since ketosis requires mobilization of fatty acids, and since increased fatty acid concentrations may decrease glucose utilization and stimulate glucose production, the in- hibition of lipolysis should reverse these. Discussion: This is a most reasonable therapeutic approach. Systems: Lipases from adipose tissue and the isolated fat cell, as well as fat pad prepara- tions, can be used for screening. Potential agents, when found, can be tested in intact animals for their ability to inhibit fatty acid release. Leads: Nicotinic acid has been used, both with only limited success in the diabetic (17). Suppression of lipolysis with nicotinic acid has recently been reported to abolish the nocturnal 352 Diabetes Mellitus rise in plasma triglyceride concentrations that occur with carbohydrate induction, but the practical application of these observations is uncertain (75). Propanolol, a beta adrenergic block, presumably acts through inhibition of epinephrine action on adipose tissue cyclic AMP dependent lipase. Propanolol may be of use in the prevention of ketoacidosis in the brittle diabetic (4). A large number of other B-adrenergic blocking agents have been prepared. 3',5' dimethylpyrazole reduces lipolysis in adipose tissue, but does not decrease fasting blood glucose concentrations, although it does reduce blood glucose concentrations in animals pretreated with glucose (42,50). A number of other compounds inhibit lipolysis, but they have not been re- ported to decrease glucose concentration (82). Antagonists to the naturally occurring fat mobilizing substance(s) (see section on Weight Reduction), if they are of importance in lipid regulation and are better characterized, may offer an additional lead. The fat mobilizing sub- stances have been reported to be diabetogenic (59). VIII. FATTY ACID OXIDATION Rationale: Blocking fatty acid oxidation results in lowered blood glucose concentrations. Discussion: This approach is supported particularly through the elucidation of the mechanism by which the ingestion of unripened ackee fruit results in hypoglycemia. Hypoglycin in the fruit is converted to an acid which inhibits many enzymatic processes and, in addition, is converted to a carnitine derivative, decreasing available free carnitine. Since carnitine is required for trans- - port of fatty acids into the mitochondria, their oxidation is depressed (21). This results in de- creased gluconeogenesis. Hypoglycin inhibits the enzyme catalyzing the oxidation of the amino acid, leucine, and the resulting accumulation of a compound, isosaluric acid, may account for the symptoms following hypoglycin's ingestion (84). The mechanism by which the decreased fatty acid oxidation leads to decreased gluconeogenesis is not certain (21,85). Systems: The biochemical steps in fatty acid oxidation have been delineated, and these offer systems for screening. Leads: The toxic effect of hypoglycin precludes its use (19). Pent-4-enoic acid has served as an analog for hypoglycin in studies of its action(21,73,85) (+)-decanoylcarnitine inhibits long-chain acylcarnitine transferase, the enzyme involved in the transport of the fatty acids, and it has been shown to inhibit hepatic ketogenesis in the isolated liver and in ketotic alloxan diabetic rats (63,89,90).When combined with insulin, it produced more of a fall in plasma ketone concentrations than with either agent alone. While the (+)-decanoylcarnitine had no effect on plasma glucose concentrations, it enhanced the hypoglycemic effect of insulin in anesthetized rats. Carnitine itself has been shown to prevent starvation ketosis in children (44). Dichloro- acetic acid lowers blood glucose concentration in diabetic animals and the mechanism of its action is evidenced by its inhibition of fatty acid oxidation by an adipose tissue preparation (80). IX. CENTRAL NERVOUS SYSTEM REGULATION Rationale: There has been for over a century evidence that the central nervous system con- tributes to the regulation of blood glucose concentration. Recently, stimulation of the ventro- medial hypothalamus has been shown to increase blood glucose and insulin glucagon concentration, but not insulin concentration (5,32) Stimulation of the premammillary but not the ventromedial area increased concentrations of plasma-free fatty acids (5). There is evidence for an insulin- sensitive center in the central nervous system (25,26,83) (see section on Weight Reduction). The stimulation of a receptor in the ventromedial hypothalamus has been postulated to activate Drugs Altering Carbohydrate and Lipid Metabolism 353 neurons which, via the sympathetic system as well as by direct neural action of the liver, partici- pate in blood glucose regulation (32). Agents which affect this system could result in lower blood glucose concentration, increased insulin secretion, and decreased lipolysis. Discussion: While the recent data indicate the central nervous system may have a major role in the regulation of glucose metabolism, the system is still too vague for the designation of specific sites within it that can be approached chemically. /Systems: Models available require either lesions in or stimuli to the central nervous system or infusions via the circulation of substances into the central nervous system, and all in the whole animal. These appear tedious to use in any screening program involving significant numbers of substances. Leads: The evidence for the role of the sympathetic system (32,43) appears to offer at present the most likely leads, examining agents which alter autonomic nervous system activities. X. POLYOL PATHWAY Rationale: If sorbitol accumulation is responsible for at least some of the complications of diabetes mellitus, then preventing its formation should prevent these complications (35). Since enhanced formation is consequent to an elevation in blood glucose concentration, this would be accomplished by returning blood glucose concentrations to normal. Alternatively, the conversion of glucose to sorbitol could be inhibited. Discussion: Whether sorbitol does play a role in the development of complications in human diabetes mellitus, and if so to what extent, is uncertain. There is also the claim that the polyol pathway is required for the stimulation of insulin release from the beta cell. Indeed, sorbitol has been reported to stimulate insulin release and an inhibitor of aldose reductase, the enzyme catalyzing the conversion of glucose to sorbitol, has been reported to inhibit glucose stimula- tion of insulin release (34). If so, then an inhibitor of sorbitol formation could decrease insulin secretion in the maturity onset diabetic and consequently increase blood glucose concentra- tion and obviate or reverse the inhibitor's potential benefits. In the juvenile diabetic, where beta cells are no longer functional, this would not be a concern. While aldose reductase from dif- ferent tissues from the same species appears to be immunologically similar, an inhibitor that is specific for tissues other than islet tissue .could be sought. Systems: Aldose reductase from several tissues has been identified, isolated, and purified. Agents inhibiting it can be tested for their ability to prevent sorbitol accumulation in vitro in lens, kidney papilla, aorta, nerve, placenta, etc., or in vivo using nerve conduction perhaps as a measure of response. , Leads: Glutaric acid derivatives have been reported to inhibit aldose reductase and in vivo to prevent cataracts and improve nerve conduction observed with galactose administration are pre- sumably due to dulcitol formation from galactose via the polyol pathway (33,66), Industrial sources are making aldose reductase inhibitors available (33,53). XI. GLYCOPROTEIN SYNTHESIS Rationale: With the beginnings of characterization of the glycoproteins composing the thickened basement membrane of the glomeruli of the kidney in the diabetic, and with evidence of increased enzymatic activity consistent with an excess of carbohydrate units in the glycoproteins (6,79) an approach through inhibition of synthesis of the glycoproteins becomes feasible. Related 354 Diabetes Mellitus to this approach is the suggestion made some years ago of an increased activity of the glucuronic acid pathway in the diabetic, a pathway by which some of the carbohydrate moieties of glycoproteins are synthesized (91). Discussion: As yet the relationship of the studies made on kidney basement membranes from other tissues is unknown, and details of the regulation of the enzymatic process are not available. An increase in the activity of thé glucuronic acid pathway has not been established, although inhibition of its activity could conceivably be effective. Systems: Methods for measuring basement membrane composition, enzymatic activities, estimat- ing the glucuronic acid pathway, and the concentrations of its intermediates in blood have been reported. Animals with vascular lesions similar to those seen in the diabetic human are being reported (30,72,92), Potential Leads: None. XII. TRIGLYCERIDE AND CHOLESTEROL SYNTHESIS Rationale and Discussion: Since the accelerated atherosclerosis manifest in the diabetic is associated with elevated concentrations of triglycerides and cholesterol, the same agents now being evaluated for the prevention or retarding of atherosclerosis in the normal individual, through their effects on triglyceride and cholesterol metabolism, should have applicability to the diabetic individuals. Systems: The same models employed in atherosclerosis studies are available, as well as digietic animals where elevated cholesterol and triglyceride exist. Leads: Clofibrate (Atromid®) and its derivatives appear to be most promising (7). Clo- fibrate may alter the progression of diabetic retinopathy (46). XIII. MULTIPLE BIOCHEMICAL PROCESSES Rationale: In addition to synthesizing and screening for agents with their design directed toward specific biochemical processes, broad screens are possible for hypoglycemic agents or agents preventing the onset of diabetes in animals known to be prediabetic or agents preventing or reversing the chronic complications of diabetes in diabetic animals. Discussion: To seek agents which do not act through insulin synthesis or secretion, animals lacking islets must serve as models. Systems: Alloxan-induced and streptozotocin-induced diabetic rats are the most readily avail- able diabetic animals. Pancreatectomized or partially pancreatectomized animals can also be used. Spontaneously diabetic animals may also be used (81). Leads: In general, most of the new chemical structures having hypoglycemic properties re- ported by the pharmaceutical industry appear to have been detected by broad screens. The compounds have been reported to decrease blood glucose concentration in either alloxan-induced or streptozotocin-induced diabetic rats, but in only a few instances has the biochemical mechanism of their action been localized. Where they are effective in animals lacking a pancreas, they are compared in terms of their prototype, the biguanides, now in clinical use, rather than the other prototype, the sulfonylureas. Pyridinium compounds are examples of such potential hypoglycemic agents (8,31). A large number of drugs with established actions other than hypoglycemia have on occasion produced hypoglycemia in man (76). A compound with hypoglycemic properties has been reported to lower urea concentrations in the diabetic (14,74) and allows increased protein intake without Drugs Altering Carbohydrate and Lipid Metabolism 355 increased urea concentrations. Elevated urea levels occur consequent to kidney damage, but such an agent would appear to offer little during the critical period prior to the onset of complica- tions, and alternate therapy is available for management of the uremia. There is no evidence the compound improves kidney function. WEIGHT REDUCTION Rationale and Discussion: Obesity will be considered in another monograph. There is no reason to believe that obesity in the diabetic is any different than in the nondiabetic, except that obesity can result in, or enhance the manifestations of, diabetes. Weight reduction's bene- fits include not only improved glucose tolerance but also decreased cardiovascular stresses. Various pharmacological approaches can be considered for achieving weight reduction. Pre- vention of intestinal absorption of food is possible, but the production of the malabsorption syndrome is an extreme which, while used in desperate situations, is unlikely to have general ap- plicability. The uncontrolled diabetic can lose weight from the disease itself, and it is possible to induce diabetes, as well as increase its severity (reversibily as with 2-deoxyglucose, diphenyl- hydantoin, etc.). Following a selected period of weight loss, diabetic control could then be in- stituted. However, the weight loss would be through glucose loss in the urine, and large quantities must be excreted to achieve weight loss with attendant discomfort, risk of urinary tract infection, loss of electrolytes, and possibly ketosis. Ketosis with its resulting anorexia could contribute significantly to the weight loss, but this would only be with all its attendant risks. Further the weight loss would be of lean body mass as well as of adipose tissue. Overall, an agent controlling appetite remains the most reasonable pharmacological approach to weight reduction. However, there is no evidence that the obese diabetic has any more of a recognition of hunger and satiety than the nondiabetic obese individual, nor that psychological problems are any less. There is the suggestion that obesity is produced through increased gluconeogenesis, and this is a stimulus to increased insulin secretion (3). This is postulated to result in the increased lipid deposition, producing a vicious cycle leading to weight gain. However, hormonal and biochemical changes thus far observed in obesity have been shown to be secondary to exogenous obesity. Appetite control in the diabetic may perhaps be different from normal. While appetite control is undoubtedly dependent on multiple stimuli to the central nervous system, the diabetic is polyphagic despite his elevated blood concentration which, in keeping with the glucostat theory of Mayer (62) would be expected to produce satiety. There is evidence that the satiety center, at least in the rat, is responsive to insulin (25, 26,83). In the insulin-lacking juvenile diabetic there would be no insulin to direct adequate glucose into the satiety center. The insulin resistance demonstrated in several tissues of obese diabetics may also exist for their satiety center. The anorexia of ketosis could then reflect perhaps in part the providing of ketone bodies as nui.rients to the satiety center. This would be in accord with the recent demonstration that ketone bodies as well as glucose can be metabolized by the brain. Much more must be done to elucidate the mechanism of appetite control, but from the evidence at hand the seeking of agents which increase glucose utilization by the satiety center seems attractive. A number of fat-mobilizing substances have been reported present in the hypothalmic-pituitary axis and in urine (59,71). A naturally occurring substance which mobilized lipid, that is, produced 356 Diabetes Mellitus lipolysis, would be expected to result in weight reduction in an individual, if the mobilized fatty acids then provided the caloric needs of the individual, and as a result his caloric intake decreased. It is reasonable to expect that with increased blood free fatty acid concentrations (and perhaps increased ketone concentrations) appetite would decrease. Systems: A satisfactory model to screen for appetite inhibitors is yet to be reported. The blood brain barrier, the time-consuming nature of electrode implants in the brain, etc., have thus far restricted its development. The models available include animals trained to eat their meals in short periods and to whom compounds are administered to see their effect on this ingestion. This is gross, and a host of toxic effects of the compounds are encountered. Gold thioglucose treated rats have their satiety center destroyed and therefore do not provide a model for study of agents acting on the center. Fat-mobilizing substances can be identified by their ability to en- hance fatty acid (or glycerol) release from adipose tissue in vitro and to increase fatty acid concentrations in vivo. Leads: Biguanides have been claimed to decrease appetite in the diabetic in the absence of the anorexia they produce. The weight loss they produce is small (2,20), and for them, as for the amphetamines and their derivatives, the duration of effect is short. The existence of the fat- mobilizing substance(s) have proved difficult to confirm and characterize. However, polypeptides from the pituitary and hypothalamus with fat-mobilizing properties have been isolated and the establishing of their structure may be near at hand. Being polypeptides, they almost certainly will be ineffective orally and their synthesis in any quantity seems unlikely. If, however, they can be shown to produce weight loss, perhaps first in experimentally obese animals, and their molecular site of action established, reasonable attempts to obtain an agent mimicking their action can be undertaken. REFERENCES: 1. Altschuld, RA, and FA Kruger 1968. Inhibition of hepatic gluconeogenesis in guinea pig by phenformin. Ann NY Acad Sci 148:612-622. 2. Appels, A, R Kotterman, H Proschek, K Hubrech, H Fierichs, HD S8ling, and W Creutzfeldt 1968. The effects of diet, tolbutamid, and buformin and their combination on body weight and several metabolic parameters in diabetes. Diabetologica 4:210-220. 3. Arky, RA, and N Freinkel 1966. Alcohol hypoglycemia. V. Alcohol infusion to test gluco- neogenesis in starvation, with special reference to obesity. New Eng J Med 274:426-433. 4. Baker, L, A Barcai, R Kaye, and N Hague 1969. Beta adrenergic blockade and juvenile diabetes: Acute studies and long-term therapeutic trial. J Pediatrics 75:19-29. 5. Barkai, A, and C Allweis 1972. Effect of electrical stimulation of the hypothalamus on plasma levels of free fatty acids and glucose in rats. Metabolism 21:921-928. 6. Beisswenger, PJ, and RG Spiro 1973. Studies on the human glomerular basement membrane. Diabetes 22:180-193. 7. Bergquist, N 1970. Serum lipids in an ambulatory diabetic clientele. Effect of therapy with atromidin (clofibrate). Acta Med Scan 187:213-218. 8. Blickens, DA, and SJ Riggi 1969. Some effects of l-methyl-4-(3-methyl-5-isoxazolyl) pyridinium chloride and phenformin on carbohydrate and lipid metabolism in mice. Diabetes 18:612-618. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. Drugs Altering Carbohydrate and Lipid Metabolism 357 Boyd, AE, III, HE Lebovits, and JB Pfeiffer 1970. Stimulation of human growth hormone secretion by L-Dopa. New Eng J Med 283:1425-1429. Brazeau, P, W Vale, R Burgus, N Ling, M Butcher, J Rivier, and R Guillemin 1973. Hypothalmic polypeptide that inhibits the secretion of immunoreactive pituitary growth hormone. Science 179:77. Brazeau, P, and R Guillemin 1974. Somatostatin: Newcomer from the hypothalamus. New Eng J Med 290:963-964. Brush, JS 1971. Purification and characterization of a protease with specificity for insulin from rat muscle. Diabetes 20:140-145. Burghen, GA, AE Kitabchi, and JE Brush 1972. Characterization of a rat liver protease with specificity for insulin. Endocrinology 91:633-642. Butterfield, WJ, BD Cox, MJ Whichelow, and GL Schless 1969. A compound which lowers blood urea in uremic diabetics. Lancet 297 (Vol 2):381-382. Cahill, GF, Jr and JS Soeldner 1974. Diabetes, glucagon, and growth hormone. New Eng J Med 291:577-578. Cafiadell, JM, A Barraquer, and CD Muiffos 1973. A new approach to the treatment of diabetic retinopathy. Diabetologica (Abs.) 9, 62. Carlson, LA 1969. Antilipolysis as a tool in the study of clinical and experimental diabetes. Diabetologica 5:361-365. Caspary, WF, and W Creutzfeldt 1973. Inhibition of intestinal amino acid transport by blood sugar lowering biguanides. Diabetologica 9:6-12. Chen, KK, WJ Fleming, and TM Lin 1961. Action of hypoglycin A on blood sugar, gastric secretion, and adipose tissue. Arch Int Phamocodyn Ther 134:435-446. Clarke, BF, and LJP Dundan 1968. Comparison of chlorpropamide and metformin treatment on the weight and blood glucose response of controlled obese diabetics. Lancet 1:123-125. Corredor, C, K Brendel, and R Bressler 1968. Studies on the mechanism of the hypoglycemic action of 4-pentenoic acid. Proc Natl Acad Sci USA 58:2299-2306. Curnow, RT, RM Carey, A Taylor, A Johanson, and F Murad 1975. Somatostatin inhibition of insulin and gastrin hypersecretion in pancreatic islet cell carcinoma. New Eng J Med 292: 1385-1386. Davidoff, F 1968. Effects of guanidine derivatives on mitochondrial function. I. Phene- thylbiguanide inhibition of respiration in mitochondria from guinea pig and rat tissues. J Clin Invest 47:2331-2343. Davidoff, F 1973. Guanidine derivatives in medicine. New Eng J Med 289:141-146. Debons, AF, I Krimsky, A From, and RJ Cloutier 1969. Rapid effects of insulin on the hypo- thalamic satiety center. Am J Physiol 217:1114-1118. Debons, AF, I Krimsky, and A From 1970. A direct action of insulin on the hypothalamic satiety center. Am J Physiol 219:938-943. Duckworth, WC, and AE Kitabchi 1973. Glucagon and insulin-degrading activity in single enzyme from skeletal muscle. Clin Res (Abs.) 21:621. Duckworth, WC, TM Chiang, EH Beachy, and AH Kang 1975. The inhibiting effect of the in vivo administration of somatostatin on platelet aggregation. The Endocrine Society 57th Annual Meetings (Abs.) 128. 358 Diabetes Mellitus 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. Dvornik, D, N Simard-Duquesne, M Krami, K Sestanj, KH Gabbay, JH Kinoshita, SD Varma, and LO Merola 1973. Polyol accumulation in galactosemic and diabetic rats: Control by an aldose reductase inhibitor. Science 182:1146-1148. Engerman, RL, MD Davis, and JMB Bloodworth Jr 1970. Retinopathy in experimental diabetes. Proc VII Cong Intl Diabetes Fed 231:261-267. Fanshawe, WJ, VJ Bauer, SR Safir, DA Blickens, and SJ Riggi 1970. Quaternary indoylpyridinium salts, oral hypoglycemic agents. J Med Chem 13:993-995. Frohman, LA, and LL Bernardis 1971. Effect of hypothalamic stimulation on plasma glucose, insulin and glucagon levels. Am J Physiol 221:1596-1603. Gabbay, KH and JH Kinoshita 1972. Mechanism of development and possible prevention of sugar cataracts. Isr J Med Sc 8:1557-1561. Gabbay, KH, and WJ Tze 1972. Inhibition of glucose-induced release of insulin by aldose reductase inhibitors. Proc. Natl Acad Sci USA 69:1435-1439. Gabbay, KH 1973. The sorbitol pathway and the complications of diabetes. New Eng J Med 288:831-836. Galloway, JA, and MA Root 1972. Newer forms of insulin. Diabetes 21 (Suppl 2):636-647. Gavin, JR, P Gorden, J Roth, JA Archer, and DN Buell 1973. Characteristics of the human lymphocyte insulin receptor. J Biol Chem 248:2202-2207. Gerich, JE, MA Charles, SR Levin, PH Forsham, and GM Grodsky 1972. In vitro inhibition of pancreatic glucagon secretion by diphenylhydantoin. J Clin Endo and Met 35:823-824. Gerich, JE, M Lorenzi, V Schneider, JH Karam, J Rivier, R Guillemin, and PH Forsham 1974. New Eng J Med 291:544-547. Gerich, JE, D Bier, P Haas, C Wood, R Byrine, and JC Penhos 1975a. In vitro and in vivo effects of somatostatin on glucose, alanine, and ketone body metabolism in the rat. The Endocrine Society 57th Annual Meeting (Abs) 128. Gerich, JE, M Lorenzi, DM Bier, V Schneider, E Tsalikian, JH Karam, and PH Forsham 1975b. Prevention of human diabetic ketoacidosis by somatostatin. New Eng J Med 292:985-989. Gerritsen, GC, and WE Dulin 1965. The effect of 5-methylpyrazole-3-carboxylic acid on carbo- hydrate and fatty acid metabolism. J Pharm Exp Ther 150:491-498. Goodner, CJ, DJ Koerker, JH Werbach, P Toivola, and CC Gale 1973. Adrenergic regulation of lipolysis and insulin secretion in the fasted baboon. Am J Physiol 224:534-539. Gravina, E, and G Gravina-Sanvitale 1969. Effect of carnitine on blood acetoacetate in fasting children. Clin Chim Acta 23:376-377. Haeckel, R, and H Haeckel 1972. Inhibition of gluconeogenesis from lactate by phenyl biguanide in the perfused guinea pig liver. Diabetologica 8:117-124. Harrold, BP, VJ Marmion, and KR Gough 1969. A double-blind controlled trail of clofibrate in the treatment of diabetic retinopathy. Diabetes 18:285-291. Hawkins, RD and H Kalant 1972. The metabolism of ethanol and its metabolic effects. Pharma- cological Reviews 24:67-157. Heath, H, WD Brigdan, TV Canever, et al. 1971. Platelet adhesiveness and aggregation in relation to diabetic retinopathy. Diabetologica 7:308-315. Hencke, WJ, and RP Doe 1967. Plasma and urinary 1ll-desoxycorticosteroid (11-DOCS) response’ to metyrapone in pituitary, renal, and hepatic disease. J Clin Endocr Metab 27:1565-1572. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. Drugs Altering Carbohydrate and Lipid Metabolism 359 Hollobaugh, SL, FA Kruger, and GJ Manwi 1967. The effect of a pyrazole derivative on plasma free fatty acids in man. Metabolism 16:996-1000. Hoppel, CL 1973. The effect of clofibrate (sodium p-chlorophenoxyisobutyrate) on mitochandrial oxidative phosphorylation. Third Pharmacology-Toxicology Program Symposium NIGMS, May 24-25, Washington, D.C. Katzen, HM, F Tietze, and D Stetten, Jr 1963. Further studies on the properties of hepatic glutathione-insulin transhydrogenase. J Biol Chem 238:1006-1011. Kinoshita, JH, D Dvornik, M Krami, and KH Gabbay 1968. The effect of an aldose reductase inhibitor on the galactose-exposed rabbit lens. Biochem Biophys Acta’158:472-475. Koerker, DJ, W Ruch, E Chideckel, J Palmer, CJ Goodner, J Ensinck, and CC Gale 1974. Somatostatin: Hypothalamic inhibitor of the endocrine pancreas. Science 184:482-483. Kreisberg, RA 1968. Kinetics of glucose utilization in obesity: The effect of phenformin. Ann NY Acad Sci 148:743-755. Kruger, FA, RA Altschuld, SL Hollobaugh, and B Jewett 1970. Phenformin inhibition of glucose transport by rat intestine. Diabetes 19:50-52. Lande, S, R Gorman, and M Bitensky 1972. Selectively blocked and des-histidine glucagons. Preparation and effects of hepatic adenylate cyclase activity. Endocrinology 90:597-604. Lockwood, DH, JJ Lipsky, F Meronk, Jr, and LE East 1971. Actions of polyamines on lipid and glucose metabolism of fat cells. Biochem Biophys Res Commun 44:600-607. Louis, LH, and JW Conn 1972. Diabetogenic polypeptide from human pituitaries similar to that excreted by proteinuric diabetic patients. Metabolism 21:1-9. Lyngsoe, J, and J Trap-Jensen 1969. Phenformin-induced hypoglycemia in normal subjects. Brit Med J 2:224-226. Lyngsoe, J, V Bitsch, and J Trap-Jensen 1972. Influence of phenformin on fat and lactate metabolism and insulin production in starved normal subjects. Metabolism 21:179-186. Mayer, J 1955. Regulation of energy intake and body weight. The glucostatic theory and the lipostatic hypothesis. Ann NY Acad Sci 63:15-93. McGarry, JD and DW Foster 1973. Acute reversal of experimental diabetic ketoacidosis in the rat with (+)-decanoylcarnitine. J Clin Invest 52:877-884. Mirsky, IA, and D Diengott 1956. Hypoglycemic action of indole-3-acetic acid by mouth in paitents with diabetes mellitus. Proc Soc Exptl Biol Med 93:105-110. Mirsky, IA, G Perisutti, and R Jinks 1957. The hypoglycemic action of metabolic derivatives of L-tryptophan by mouth. Endocrinology 60:318-324. Morrison, AD, and AI Winegrad 1973. Inhibition of polyol pathway ameliorates effects of elevated glucose levels in aorta. Clin Res (Abs) 21:632. O'Brien, JR 1968. Effects of salicylates on human platelets. Lancet 1:779-783. Patel, YC, GC Weir, and S Peichlin 1975. Anatomic distribution of somatostatin (SRIF) in brain and pancreatic islets as studied by radioimmunoassay (RIA). The Endocrine Society 57th Annual Meeting (Abs) 127. Pelletier, G, R Leclerc, and Dubé 1975. Immunohistochemical localization of somatostatin in in the rat brain. The Endocrine Society 57th Annual Meeting (Abs) 127. Powell, EDU, and RA Field 1965. Salicylates and diabetic retinopathy. Diabetes (Abs) 14:462. Redding, TW, and AV Schally 1972. . Effect of hypothalamic preparations on human omental adipose tissue in vitro. Metabolism 21:499-506. 360 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. Diabetes Mellitus Renold, AE, DP Cameron, M Amherdt, W Stauffacher, E Marliss, L Orci, and C Rouiller 1972. Endocrine-metabolic anomalies in rodents with hyperglycemic syndromes of hereditary and/or environmental origin in Impact of Insulin on Metabolic Pathways. A Symposium, Jerusalem, Oct 1971. E Shafrir, Ed, Academic Press, NY, pp 15-32. Ruderman, NB, CJ Toews, C Lowy, I Vreeland, and E Shafrir 1970. Inhibition of hepatic gluconeogenesis and fatty acid oxidation by pent-4-enoic acid. Am J Physiol 219:51-57. Schless, GL, WJ Butterfield, BD Cox, and MJ Whichelow 1970. 5 3 ! Je | : A ) . & ome nr ad pmin . : | siaec 4 . Loh . Eo ASE = i | ¥ 2 13 b iin Nr a - a | 3 WRF ceil we el bari x i - : MPR mt ll mis SBN Hani. i ¥ YF g ga 2 Ls Rl opie Send 49 fer | y y RE eo RAD a nm ee iti et En To prem =r re rs E i hol A gem Ba oo Fe gS 1 Fe § DHEW Publication (NIH) ° re (N51 3 C03174331y